Model family by OpenAI

GPT Terra

OpenAI's fast, efficient GPT tier for visual reasoning, computer use, and writing.

Where it sits in OpenAI's lineup

GPT TerraOpenAI's fast, efficient GPT tier for visual reasoning, computer use, and writing.This model
GPT SolOpenAI's GPT model tier for demanding everyday and agentic work.GPT-6.1 Sol
GPT LunaOpenAI's most efficient GPT model tier for focused, high-volume tasks.GPT-6 Luna
GPT AstraOpenAI's highest-capability GPT model tier.GPT-6 Astra
GPT-5OpenAI's GPT-5 frontier models, released before the Sol, Terra and Luna tiers.GPT-5.5
gpt-ossOpenAI's open-weight GPT models under the Apache 2.0 license.gpt-oss-120b
GPT CodexOpenAI's GPT models tuned for agentic coding in Codex.GPT-5.3-Codex

Made by: OpenAI
Releases: 2 tracked
Current release: GPT-6.1 Terra

Strengths and weaknesses

Compare with other models Up to 3 beside GPT-6.1 Terra

GPT-6.1 Terra

Suggested comparisons

GPT-6.1 Terra is the newest GPT Terra release. Add its earlier releases to see how the line moved.

Search models by name, maker or family

GPT-6.1 TerraOpenAIThis page, always on the chart

Among the 40 current releases on this chart, GPT-6.1 Terra ranks 4th of 28 on Math, 5th of 24 on Multimodal and 8th of 31 on Agentic coding. On Computer use, OpenAI's own figures on OSWorld 2.1 put GPT-6.1 Terra 6.7 points under GPT-6 Astra, 3.8 points under GPT-6.1 Sol and 18.7 points above GPT-6 Luna. Answers in full: not yet published by OpenAI or any independent evaluator, checked Nov 2, 2026.

Each axis is one public benchmark at one setting, on its own fixed 0 to 100 scale, so adding a model never moves another. A dotted bridge crosses a missing result, which is never a zero. A hollow mark is a result from another source: the maker's own figure or another lab's run, where the benchmark's usual source has not tested the model yet. The readout names it.

On the chart

GPT-6.1 TerraThis pageResults on 11 of 12 axes$1 / $5For prompts under 272K tokens; $2 / $7.50 at 272K or more, as OpenRouter lists it.97 tokens/sFirst token in 275 s

Prices and speeds read between Oct 26, 2026 and Nov 2, 2026.

GPT-6.1 Terra on each axis11 of 12 with a result

The share of Terminal-Bench 4.0's long terminal tasks the model finishes on its own: services, kernels, proofs and data pipelines, graded on the final artifact with no partial credit.
Terminal-Bench 4.0, pass@1 averaged over 3 runs, mini-swe-agent harness
- GPT-6.1 Terra49.5%
The share of SciCode's scientific programming subproblems the model's code solves on the first attempt.
SciCode, subproblem accuracy, pass@1, highest effort run
- GPT-6.1 Terra55.1%
The share of FrontierMath's held-out research-grade problems (tiers 1 to 3) the model answers correctly.
FrontierMath Tiers 1-3, accuracy, pass@1, highest effort run
- GPT-6.1 Terra90.9%
The share of GPQA Diamond's graduate-level biology, chemistry and physics questions the model answers correctly.
GPQA Diamond, accuracy, pass@1, highest effort run
- GPT-6.1 Terra94.3%
The share of Humanity's Last Exam's 2,158 text-only questions, expert-written at the edge of human knowledge, the model answers correctly without tools.
Humanity's Last Exam, accuracy, text-only questions, no tools
- GPT-6.1 Terra46.2%
The share of each AutomationBench task's objectives the model completes by calling REST APIs across simulated business apps (Gmail, Sheets, Slack, Salesforce, Zendesk, Jira, HubSpot), averaged over tasks. A task that breaks one of its policy guardrails scores zero.
AutomationBench-AA, score, Artificial Analysis run
- GPT-6.1 Terra60.2%
How much of each OSWorld 2.1 workflow the model completes by operating a real Ubuntu desktop through screenshots, mouse and keyboard: all 108 long tasks across apps, files, mail and browsers, on the September 2026 v2.1 release of the task files. A point is the mean share of each task's weighted checks the model passes (the partial score, as a percentage). The OSWorld team's own runs come first. Where the team has not run a release, the point is another lab's run or the maker's own published figure, marked with who ran it and the benchmark it ran: OSWorld 2.1 in the maker's harness, OSWorld 2.0's offline set or the earlier OSWorld-Verified. Those points are drawn so the release has a figure, and they are not ranked against the team's runs.
OSWorld 2.1, partial score, full 108-task set, v2.1 release. Where it has no run of a model, the maker's own figure or another lab's run stands in, marked with its benchmark.
- GPT-6.1 Terra74.2%reported by OpenAIBenchmark: OSWorld 2.1, score, offline set, v2.1 release, where the axis reads partial score, full 108-task set. From another source, so it is drawn and marked but not ranked with the others.OpenAI's own run at maximum reasoning effort, from its launch chart, at $1.19 per task. OpenAI does not say whether the score counts partial credit.
The share of AA-LCR's questions the model answers correctly by reasoning across long sets of real documents, about 100K tokens each.
AA-LCR v1.1, accuracy, Artificial Analysis run
- GPT-6.1 Terra82.7%
The share of MMMU-Pro's college-level questions over charts, diagrams and photos the model answers correctly.
MMMU-Pro, overall accuracy, no tools
- GPT-6.1 Terra84.7%
The share of AA-Omniscience's 6,000 knowledge questions across 42 economically relevant topics the model answers correctly from memory, without search.
AA-Omniscience Accuracy, accuracy, Artificial Analysis run
- GPT-6.1 Terra47.6%Hallucination rate 34.5% in the same run (lower is better)
The share of SpeechMap's 2,120 sensitive and controversial but legal requests the model answers in full: arguments, essays and explanations on contested themes in politics, religion, race, sex, history and similar topics, four wordings per theme across 530 themes. SpeechMap's judge rates each response complete, evasive (hedged or redirected), a denial, or an error (the provider's API blocked the request). Only complete answers count, so the point is the share answered in full: 100 minus the share hedged, refused or blocked by the provider. Higher means more lawful requests answered in full. The share blocked at the provider is printed apart in the table, because an API error is not the model declining. Where SpeechMap runs a release both with reasoning off and with reasoning on (a -reasoning or -thinking entry), the reasoning-on entry is used; a release with one entry uses that entry.
SpeechMap, % Complete over 2,120 contested requests, reasoning-on entry
- GPT-6.1 TerraNot yet published by OpenAI or any independent evaluator, checked Nov 2, 2026
The share of MultiChallenge's multi-turn conversations in which the model keeps every instruction it was given earlier in the conversation.
MultiChallenge, accuracy, Scale AI run
- GPT-6.1 Terra74.2%On Oct 28, 2026, Scale AI ranked it 6th, tied with Muse Spark Open.Plus or minus 5.4 points (95% interval)

Beside the benchmarks

Cost, speed, capacity, quality and accuracy for GPT-6.1 Terra, with where each figure sits among the 40 models GILD tracks. These measures stand beside the chart and never feed it.

Cost

$0.96per task

Cost per task. One Intelligence Index task at list price, every token counted. Run at max effort.

Cached input: $0.05 per 1M tokens

Speed

91tokens a second

Output speed. How fast it writes once the answer starts. Run at max effort.

First token: 78 s
Whole answer: 83 s

Capacity

1.05Mtokens

Context window. The most one request can hold, as the maker states it.

Max output: 128K
Reads: Text, Images

Quality

51.0of 100

Intelligence Index. One composite score across independent evaluations. Run at max effort.

GDPval knowledge work: 1,476 Elo
Terminal-Bench 4.0: 55.6%

Hallucination

34.5%wrong when it should say “I don’t know”

Hallucination rate. Of the questions it can't get right, how often it answers anyway. Lower is better. Run at max effort.

Knowledge accuracy: 47.6%

Where you can get it

7of 19 routes

Routes offering it. Of the clouds, gateways and coding products GILD tracks, those offering it today.

Open weights: No
Licence: OpenAI Services Agreement

Every route and price

Figures from Artificial Analysis, OpenAI and GILD's route matrix, read Oct 27 to Nov 2, 2026. Each measure's definition and source open under . Not yet published for GPT-6.1 Terra: Epoch Capabilities Index.

Every value with its benchmark, run and source

GPT-6.1 Terra's result on each axis, with the benchmark, the setting and run, who reported it, the source and the date
Axis	Benchmark and setting	GPT-6.1 Terra	Run	Source	Date
Agentic coding	Terminal-Bench 4.0Pass@1 averaged over 3 runs, mini-swe-agent harness	49.5%	GPT-6.1 TerraEffort: not published. Harness: mini-swe-agent.	Vals.ai	Oct 26, 2026
Code	SciCodeSubproblem accuracy, pass@1, highest effort run	55.1%	GPT-6.1 Terra (Max)Effort: max. Harness: none.	Artificial Analysis	Nov 2, 2026
Math	FrontierMath Tiers 1-3Accuracy, pass@1, highest effort run	90.9%	Effort: max. Harness: none.	Epoch AI	Oct 26, 2026
Science	GPQA DiamondAccuracy, pass@1, highest effort run	94.3%	Effort: max. Harness: none.	Epoch AI	Oct 26, 2026
Hard reasoning	Humanity's Last ExamAccuracy, text-only questions, no tools	46.2%	GPT-6.1 Terra (Max)Effort: max. Harness: none.	Artificial Analysis	Nov 2, 2026
Tool use	AutomationBench-AAScore, Artificial Analysis run	60.2%	GPT-6.1 Terra (Max)Effort: max. Harness: Artificial Analysis AutomationBench harness.	Artificial Analysis	Nov 2, 2026
Computer use	OSWorld 2.1Score, offline set, v2.1 release. OpenAI's own run at maximum reasoning effort, from its launch chart, at $1.19 per task. OpenAI does not say whether the score counts partial credit.Stands in for OSWorld 2.1, partial score, full 108-task set, v2.1 release, which has no run of GPT-6.1 Terra yet.	74.2%	GPT-6.1 Terra (Max)Effort: max. Harness: OpenAI computer-use harness.	OpenAIReported by OpenAI	Oct 26, 2026
Long context	AA-LCR v1.1Accuracy, Artificial Analysis run	82.7%	GPT-6.1 Terra (Max)Effort: max. Harness: none.	Artificial Analysis	Nov 2, 2026
Multimodal	MMMU-ProOverall accuracy, no tools	84.7%	GPT-6.1 Terra (Max)Effort: max. Harness: none.	Artificial Analysis	Nov 2, 2026
Factuality	AA-Omniscience AccuracyAccuracy, Artificial Analysis run	47.6%Hallucination rate 34.5% in the same run (lower is better)	GPT-6.1 Terra (Max)Effort: max. Harness: none.	Artificial Analysis	Nov 2, 2026
Answers in full	SpeechMap% Complete over 2,120 contested requests, reasoning-on entry	Not yet published by OpenAI or any independent evaluator, checked Nov 2, 2026
Instruction following	MultiChallengeAccuracy, Scale AI run. On Oct 28, 2026, Scale AI ranked it 6th, tied with Muse Spark Open.	74.2%Plus or minus 5.4 points (95% interval)	GPT 6.1 TerraEffort: not published. Harness: none.	Scale AI SEAL	Oct 28, 2026

Price and speed

GPT-6.1 Terra lists at $1 per million input tokens and $5 per million output tokens. For prompts under 272K tokens; $2 / $7.50 at 272K or more, as OpenRouter lists it. OpenAI pricing, Oct 26, 2026.
GPT-6.1 Terra, at max effort, writes 97.2 output tokens per second on OpenAI's API, and its first token arrives 275 s after the request. Artificial Analysis, Nov 2, 2026.
Fastest on the chart is the shortest time to a 500-token answer: time to first token plus 500 tokens at the median output speed.

Every model with a result, axis by axis

57 models. A maker's own figure and another lab's run are marked with their benchmark; a missing result reads "Not yet published", never zero.

The result of every model GILD has benchmark results for, on each of the 12 axes. A vendor-reported value is marked. A model's run notes open under its row from its name, each with the axes it covers.
Model	Agentic coding	Code	Math	Science	Hard reasoning	Tool use	Computer use	Long context	Multimodal	Factuality	Answers in full	Instructions
GPT-6.1 Terra	49.5%	55.1%	90.9%	94.3%	46.2%	60.2%	74.2%Reported by OpenAIOSWorld 2.1	82.7%	84.7%	47.6%	Not yet published	74.2%
Computer use. OpenAI's own run at maximum reasoning effort, from its launch chart, at $1.19 per task. OpenAI does not say whether the score counts partial credit. OpenAI Instructions. On Oct 28, 2026, Scale AI ranked it 6th, tied with Muse Spark Open. Scale AI SEAL
GPT-6.1 Sol	55.1%	54.2%	93.7%	95.4%	52.9%	64.9%	78%Reported by OpenAIOSWorld 2.1	83%	86%	61.8%	69.5%	82%
Computer use. OpenAI's own run at maximum reasoning effort, from the GPT-6.1 Terra launch chart, at $1.15 per task against $8.17 for GPT-6 Astra. OpenAI does not say whether it counts partial credit. OpenAI Answers in full. SpeechMap tested it through OpenRouter at medium reasoning and counted 21.7% of requests denied and 8.8% evaded. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 1st. Scale AI SEAL
GPT-6 Luna	13.6%	54.6%	78.9%	90.5%	38.5%	53.2%	48.9%Run by AnthropicOSWorld 2.1	83.3%	79.7%	39.6%	63.4%	Not yet published
Computer use. Anthropic's run of GPT-6 Luna through OpenAI's API with OpenAI's context compaction, 500 steps, averaged over five attempts, from the Claude Haiku 5.5 system card. Anthropic Answers in full. SpeechMap tested it through OpenRouter at medium reasoning and counted 27.4% of requests denied and 9.1% evaded. SpeechMap.AI
GPT-6 Astra	59.6%	56.5%	93.7%	95.8%	54.7%	68.5%	75.8%Run by Meta	80.7%	86.9%	61.6%	68.2%	Not yet published
Agentic coding. Vals.ai ranked it 4th on its leaderboard of Oct 24, 2026. Vals.ai Computer use. Meta's run of GPT-6 Astra at max effort in its own test framework on all 108 tasks: the average partial score. It fully completed 34.3% of tasks. Meta Answers in full. Reasoning cannot be turned off on this model. SpeechMap counted 21.7% of requests denied and 10.0% evaded. SpeechMap.AI
GPT-5.6 Sol	37.9%	57.1%	89.1%	93.5%	49.5%	60.1%	62.7%Run by OSWorld teamOSWorld 2.0	84%	83.4%	69.7%Run by Epoch AISimpleQA Verified	Not yet published	72.7%Run by Artificial AnalysisIFBench
Computer use. OSWorld team's official run of GPT-5.6 Sol at max reasoning, batch tool, 500 steps. OSWorld team Factuality. Epoch AI's run at max effort. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Instructions. Artificial Analysis ran GPT-5.6 Sol at its highest published effort in its own evaluation harness. Artificial Analysis
GPT-5.6 Terra	22.7%	55%	86%	93.3%	42.9%	59.6%	Not yet published	83%	80.7%	42.3%	Not yet published	71.2%Run by Artificial AnalysisIFBench
Instructions. Artificial Analysis ran GPT-5.6 Terra at its highest published effort in its own evaluation harness. Artificial Analysis
GPT-5.6 Luna	11.6%	53.6%	82.1%	91.6%	39.5%	50.2%	Not yet published	83.7%	78.6%	41%Run by Epoch AISimpleQA Verified	Not yet published	Not yet published
Factuality. Epoch AI's run at max effort. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI
GPT-5.5	14.6%Run by Artificial AnalysisTerminal-Bench 4.0	55.8%	85.3%	93.5%Run by Artificial Analysis	45.8%	47.3%	49.5%Run by OSWorld teamOSWorld 2.0	84.3%	79.9%	63%Run by Epoch AISimpleQA Verified	Not yet published	75.9%Run by Artificial AnalysisIFBench
Agentic coding, Science and Instructions. Artificial Analysis ran GPT-5.5 at its highest published effort in its own evaluation harness. Artificial Analysis Computer use. The OSWorld team's own run of GPT-5.5 at xhigh reasoning with batched tool calls and a 500-step budget, on the first OSWorld 2.0 task release. OSWorld team Factuality. Epoch AI's run at xhigh effort, the highest it ran. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI
GPT-5.4	Not yet published	Not yet published	78.6%	93.3%	43.7%	Not yet published	Not yet published	82%	78.4%	Not yet published	Not yet published	73.9%Run by Artificial AnalysisIFBench
Instructions. Artificial Analysis ran GPT-5.4 at its highest published effort in its own evaluation harness. Artificial Analysis
gpt-oss-120b	14.2%Run by Terminal-Bench (via Epoch AI)Terminal-Bench 2.0	34%	88.9%Run by Epoch AIOTIS Mock AIME 2024-2025	75.8%	19.6%	0.2%	Not yet published	52%	Not yet published	16.8%Reported by OpenAISimpleQA	36.3%	45.3%
Agentic coding. The Terminal-Bench team's run on its version 2.0 leaderboard with the mini-swe-agent harness, the same agent the axis uses. Terminal-Bench 2.0 is the earlier, easier version of the test. Terminal-Bench (via Epoch AI) Math. Epoch AI's own run at high reasoning, the highest gpt-oss-120b offers. Epoch has not run it on FrontierMath; this competition-math set is easier. Epoch AI Tool use. Artificial Analysis publishes 0.2 for its high-effort run on this test, far below the model's other results. Artificial Analysis Factuality. OpenAI's own run in its model card, without browsing. It answered 78.2% of questions wrongly; OpenAI counts that share of all questions as its hallucination rate. OpenAI Answers in full. SpeechMap counted 63.3% of requests denied and 0.3% evaded. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 31st. Scale AI SEAL
GPT-6 Sol	44.4%	57.6%	89.8%	94.3%	47.9%	61.6%	64.4%Reported by OpenAIOSWorld 2.0	83.7%	82.9%	54.5%	64%	Not yet published
Computer use. OpenAI's own run at maximum reasoning effort, from its launch chart, at $3.25 per task. OpenAI Answers in full. SpeechMap tested it through OpenRouter at medium reasoning and counted 22.7% of requests denied and 13.2% evaded. SpeechMap.AI
GPT-5.3-Codex	Not yet published	Not yet published	Not yet published	91.5%Run by Artificial Analysis	42.5%	Not yet published	Not yet published	83.3%	78.5%	Not yet published	Not yet published	75.4%Run by Artificial AnalysisIFBench
Science. Artificial Analysis ran GPT-5.3-Codex at its highest published effort in its own evaluation harness; Epoch AI has no qualifying run. Artificial Analysis Instructions. Artificial Analysis ran GPT-5.3 Codex at its highest published effort in its own evaluation harness. Artificial Analysis
Claude Haiku 5.5	36.4%	55%	75.1%	89.6%	44.4%	35.4%	72.4%Reported by AnthropicOSWorld 2.1	82.7%	Not yet published	34.7%	78.8%	67.7%
Computer use. Anthropic's own run at max effort in its computer-use harness, averaged over five attempts, on the 82 tasks that need no internet. It fully completed 37.1% of them. Anthropic Answers in full. SpeechMap tested it at medium thinking and counted 18.3% of requests denied and 2.5% evaded. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 10th. Scale does not say which reasoning setting it tested. Scale AI SEAL
Claude Sonnet 5.5	64.1%	61%	88.8%	95.6%	55%	71.8%	80.1%Reported by Anthropic	82.7%	Not yet published	51.5%	81.6%	Not yet published
Code, Hard reasoning, Tool use, Long context and Factuality. Artificial Analysis tested Sonnet 5.5 as Anthropic serves it by default, with its safeguard fallback model on. Artificial Analysis Computer use. Anthropic's own run at max effort in its computer-use harness, averaged over five runs. It fully completed 43.5% of the tasks. Anthropic Answers in full. SpeechMap tested it through OpenRouter at medium thinking and counted 13.5% of requests denied and 4.6% evaded. SpeechMap.AI
Claude Opus 5.5	65.2%	66.9%	91.2%	90.6%	61.4%	69.5%	81.8%Reported by Anthropic	84.7%	87.7%	65.5%	82.8%	Not yet published
Agentic coding. Vals.ai ranked it 1st on its leaderboard of Oct 24, 2026. Vals.ai Code, Hard reasoning, Tool use, Long context, Multimodal and Factuality. Artificial Analysis tested Opus 5.5 as Anthropic serves it by default, with its safeguard fallback model on. Artificial Analysis Computer use. Anthropic's own run at max effort in its computer-use harness, averaged over five runs. It fully completed 48.7% of the tasks. Anthropic Answers in full. SpeechMap tested it through OpenRouter at medium thinking and counted 12.7% of requests denied and 2.2% evaded. Another 2.3% failed at the provider and count against the score. SpeechMap.AI
Claude Fable 5.1	58.1%	63.1%	90.2%	93.4%Run by Vals.ai	59.1%	59.4%	80.7%Reported by Anthropic	85.3%	86.3%Run by Moonshot AI	68.4%	72.4%	Not yet published
Code, Hard reasoning, Tool use, Long context and Factuality. Artificial Analysis tested Fable 5.1 as Anthropic serves it by default, with its safeguard fallback model on. Artificial Analysis Science. Vals.ai's run at max effort, ranked 9th on its GPQA Diamond board, last updated September 1, 2026, before Vals archived the board. Vals.ai Computer use. Anthropic's own run on the v2.1 tasks, averaged over five runs. It replaces the 77.9% in the Fable 5.1 system card, which used the August 2026 task set. Anthropic Multimodal. Moonshot AI's run at max effort with Anthropic's fallback on, from the Kimi K3.1 Pro launch table, by the official MMMU-Pro protocol. Artificial Analysis has not run it. Moonshot AI Answers in full. With reasoning on, SpeechMap counted 15.1% of requests denied. Another 11.1% failed at the provider and count against the score, far more than any other Claude release. SpeechMap.AI
Claude Opus 5	53.5%	56.4%	85.6%	93.9%	54.9%	56.6%	77.7%	79.3%	84.7%	60.9%	73.7%	Not yet published
Computer use. The OSWorld team's own run at max effort, with its batch tool and a 500-step budget. It fully completed 44.33% of tasks. OSWorld team Answers in full. With reasoning on, SpeechMap counted 22.3% of requests denied and 3.5% evaded. With reasoning off, the model completes 59.1%. SpeechMap.AI
Claude Sonnet 5	10.6%	54.3%	65.6%	80.3%	41.3%	36.5%	57%Reported by Anthropic	82%	77.3%	40.1%	Not yet published	Not yet published
Computer use. Anthropic reran it for the Sonnet 5.5 launch, averaged over five runs. The share of tasks fully completed is 25.6%. Anthropic
Claude Fable 5	41.4%	61%	87%	85.9%	55.5%	54.1%	Not yet published	82.3%	Not yet published	65.4%	64.8%	63.5%Run by Artificial AnalysisIFBench
Code, Hard reasoning, Tool use, Long context and Factuality. Artificial Analysis tested Fable 5 with its Opus 4.8 fallback switched on, the setup its model page lists. Artificial Analysis Answers in full. With reasoning on, SpeechMap counted 27.9% of requests denied and 5.0% evaded. Another 2.4% failed at the provider and count against the score. SpeechMap.AI Instructions. Artificial Analysis ran Claude Fable 5 at max effort in its own evaluation harness. The run routes some requests to Opus 4.8 as a fallback. Artificial Analysis
Claude Opus 4.8	23.2%	54.4%	80%	91%	48.7%	45.6%	54.8%Run by OSWorld teamOSWorld 2.0	77.7%	Not yet published	48.8%	67.4%	62.2%Run by Artificial AnalysisIFBench
Computer use. The OSWorld team's own run at max effort with its batched tool and a 500-step budget, on the v2026.08.08 task release. It fully completed 20.6% of tasks; the standard tool setting scored 49.33. OSWorld team Answers in full. With thinking on, SpeechMap counted 25.2% of requests denied and 7.0% evaded. With thinking off, the model completes 43.4%. SpeechMap.AI Instructions. Artificial Analysis ran Claude Opus 4.8 at max effort in its own evaluation harness. Artificial Analysis
Claude Opus 4.7	Not yet published	54.5%	70.2%	86.4%	42.3%	Not yet published	48.9%Run by OSWorld teamOSWorld 2.0	78.7%	78.8%	48.9%	70.2%	58.6%Run by Artificial AnalysisIFBench
Computer use. The OSWorld team's own run at max effort with its batched tool and a 500-step budget, on the v2026.08.08 task release. It fully completed 18.2% of tasks; the standard tool setting scored 49.1. OSWorld team Answers in full. With reasoning on, SpeechMap counted 22.0% of requests denied and 5.1% evaded. With reasoning off, the model completes 71.5%. SpeechMap.AI Instructions. Artificial Analysis ran Claude Opus 4.7 at max effort in its own evaluation harness. Artificial Analysis
Claude Sonnet 4.6	3%Run by Artificial AnalysisTerminal-Bench 4.0	50.1%	71.1%Run by Epoch AIOTIS Mock AIME 2024-2025	78.8%	33.6%	20.1%	41.5%Run by OSWorld teamOSWorld 2.0	80%	73.3%	40.9%	35.8%	56.6%Run by Artificial AnalysisIFBench
Agentic coding. Artificial Analysis's run of the October 29 build at high effort in its own agent harness. Vals.ai has run only the October 6 preview build. Artificial Analysis Agentic coding. Unusual. Artificial Analysis's run of the October 29 build at high effort in its own agent harness. Vals.ai has run only the October 6 preview build. Oct 28, 2026 Artificial Analysis Math. Epoch AI's run at max effort, the highest effort it ran. Its 32K thinking budget run scored 85.8, and high 75.6. Epoch AI Computer use. The OSWorld team's own run at max effort with a 500-step budget, on the v2026.08.08 task release. It fully completed 8.3% of tasks; its medium-effort run scored 33.9. OSWorld team Answers in full. With thinking on, SpeechMap counted 45.4% of requests denied and 18.3% evaded. With thinking off, the model completes 29.1%. SpeechMap.AI Instructions. Artificial Analysis ran Claude Sonnet 4.6 at max effort in its own evaluation harness. Artificial Analysis
Claude Opus 4.6	Not yet published	Not yet published	66%	88.4%	39.9%	Not yet published	Not yet published	78%	75.4%	47%	59.9%	37.2%
Answers in full. With thinking on, SpeechMap counted 27.9% of requests denied and 10.4% evaded. With thinking off, the model completes 49.8%. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked this run 35th. The same model with thinking off scores 56.02 on Scale's board. Scale AI SEAL
Claude Opus 4.5	Not yet published	Not yet published	34.4%	86%	30.1%	Not yet published	Not yet published	77.3%	74%	46.6%	57.5%	59%
Math. Epoch AI's run with a 32K-token thinking budget. Epoch AI Science. Epoch AI's run with a 32K-token thinking budget. With thinking off it scores 80.7%. Epoch AI Answers in full. With extended thinking on, SpeechMap counted 29.1% of requests denied and 13.3% evaded. With it off, the model completes 50.8%. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 18th, with extended thinking on. Scale AI SEAL
Claude Haiku 4.5	Not yet published	42.2%	66.7%Run by Epoch AIOTIS Mock AIME 2024-2025	71.2%	10.4%	3.2%	Not yet published	74.3%	58.6%	18%	24.7%	50.5%
Math. Epoch AI's run with a 32K thinking budget, the highest it ran. Without extended thinking the same model scored 35.8. Epoch AI Science. Epoch AI's run with a 32K-token thinking budget. With thinking off it scores 60.5%. Epoch AI Tool use. Artificial Analysis publishes 3.2 for the reasoning version of Haiku 4.5 on this test, far below the model's other results. Artificial Analysis Tool use. Unusual. Artificial Analysis publishes 3.2 for the reasoning version of Haiku 4.5 on this test, far below the model's other results. Oct 28, 2026 Artificial Analysis Answers in full. With extended thinking on, SpeechMap counted 39.0% of requests denied and 36.4% evaded. With it off, the model completes 20.8%. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 29th, with extended thinking on. Scale AI SEAL
Claude Sonnet 4.5	Not yet published	45.7%	23.9%	82.3%	17.8%	14.1%	Not yet published	72.3%	68.7%	32.9%	31.4%	55.3%
Math. Epoch AI's run with a 32K-token thinking budget. Epoch AI Science. Epoch AI's run with a 59K-token thinking budget. With thinking off it scores 73.7. Epoch AI Answers in full. With reasoning on, SpeechMap counted 48.3% of requests denied and 20.2% evaded. With reasoning off, the model completes 32.3%. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 25th, with extended thinking on. Scale AI SEAL
Gemini 4 Argon	57.6%	61.8%	86%	96.4%	57.1%	77.5%	69.2%Reported by GoogleOSWorld 2.0	79.7%	Not yet published	48.3%	87%	78.7%
Computer use. Google's own run at its highest thinking setting in its Gemini computer-use harness, best of 3 runs. No outside lab has run it. Google Answers in full. SpeechMap tested it through OpenRouter at medium reasoning and counted 10.2% of requests denied and 2.4% evaded. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 2nd. Scale AI SEAL
Gemini 3.8 Flash	19.7%	56.6%	68.4%	95.4%	47.8%	59.9%	59%Reported by GoogleOSWorld 2.0	81.3%	85.6%	50.5%	86.1%	Not yet published
Computer use. Google's own run in its Gemini computer-use harness, best of 3 runs, on the task files before the August patch. Google does not print the thinking level. Google Answers in full. SpeechMap tested it through OpenRouter with reasoning on and counted 11.7% of requests denied and 1.8% evaded. SpeechMap.AI
Gemma 4 31B	Not yet published	45.5%	Not yet published	85.7%Run by Artificial AnalysisGPQA Diamond	23.6%	4.9%	Not yet published	69.7%	73.4%	10.4%Run by Epoch AISimpleQA Verified	82.3%	75.6%Run by Artificial AnalysisIFBench
Science. Artificial Analysis's run of the reasoning version. Epoch AI, the axis's source, ran it only at minimal effort, which is not drawn. Artificial Analysis Tool use. Artificial Analysis publishes 4.9 for the reasoning version of Gemma 4 31B on this test, far below its other results. Artificial Analysis Tool use. Unusual. Artificial Analysis publishes 4.9 for the reasoning version of Gemma 4 31B on this test, far below its other results. Oct 28, 2026 Artificial Analysis Factuality. Epoch AI ran SimpleQA Verified on the instruction-tuned release. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Answers in full. SpeechMap counted 16.8% of requests denied and 0.9% evaded. SpeechMap.AI Instructions. Artificial Analysis's run of the reasoning version on IFBench, a test of precise output constraints. Scale AI has not run it on MultiChallenge. Artificial Analysis
Gemini 3.1 Pro Preview	2.5%	58.7%	59.6%	94.4%	47%	35.4%	Not yet published	82%	82.4%	54.9%	59.9%	71.4%
Agentic coding. Vals.ai tested the February 2026 preview build. Vals.ai Agentic coding. Unusual. Vals.ai tested the February 2026 preview build. Oct 24, 2026 Vals.ai Long context. Artificial Analysis tested the February 19, 2026 preview build. Artificial Analysis Answers in full. With reasoning on, SpeechMap counted 30.9% of requests denied and 8.8% evaded. SpeechMap.AI Instructions. On Oct 28, 2026, Scale AI ranked it 8th. Scale does not say which reasoning setting it tested. Scale AI SEAL
Gemini 3.6 Flash	7.1%Run by Artificial AnalysisTerminal-Bench 4.0	53.4%	58.9%	94.1%	40.8%	53%	83%Reported by GoogleOSWorld-Verified	80%	83.2%	66.2%Run by Epoch AISimpleQA Verified	87.2%	Not yet published
Agentic coding. Artificial Analysis ran Terminal-Bench 4.0 in its own agent harness at high effort. Artificial Analysis Computer use. Google's own figure from the Gemini 3.6 Flash model card, results as of July 2026. It does not state the harness, step budget or thinking level. No independent lab has run it. Google Factuality. Epoch AI's run at high effort, the highest it ran. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Answers in full. SpeechMap tested Gemini 3.6 Flash through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. SpeechMap.AI
DeepSeek V4.1 Flash	31.8%	55.9%	71.2%	93.2%	43.3%	66.6%	69.3%Run by Moonshot AIOSWorld 2.1	82.7%	80.6%	39.6%	Not yet published	74.9%
Agentic coding. Vals.ai ranked the October 17 build 19th on its leaderboard of Oct 24, 2026. Vals.ai Computer use. Moonshot AI's run of the October 17 build at max effort in its Kimi Agent harness, from the Kimi K3.1 Pro launch table. Moonshot does not name the metric. Moonshot AI Instructions. On Oct 28, 2026, Scale AI ranked the October 17 build 5th. Scale AI SEAL
DeepSeek V4 Pro	14.1%	51%	64.6%	91.7%	41%	56.7%	Not yet published	80.3%	Not yet published	45.2%	72%	Not yet published
Agentic coding. Vals.ai ranked it 38th of 58 on its leaderboard of Oct 29, 2026. Vals.ai Factuality. Of the questions it did not know, it answered 94.1% wrongly when it should have declined. Artificial Analysis prints this hallucination rate beside accuracy. Artificial Analysis Answers in full. SpeechMap counted 27.6% of requests denied and 0.3% evaded. SpeechMap.AI
Kimi K3.1 Pro	63.6%	60.8%	Not yet published	94.2%Reported by Moonshot AI	52.5%	73.3%	79.6%Reported by Moonshot AIOSWorld 2.1	86.3%	83.9%	48.9%	Not yet published	Not yet published
Science. Moonshot AI's own figure, at its highest reasoning setting with no tools. No independent lab has run it yet. Epoch AI scores the earlier Kimi K3 at 93.1. Moonshot AI Computer use. Moonshot AI's own run at max effort in its Kimi Agent harness. Moonshot does not name the metric or task set; its other rows reprint Anthropic's full-set partial scores. Moonshot AI
Kimi K3	17.2%	59.5%	72.2%	93.1%	46.9%	58.3%	Not yet published	88.7%	80.5%	50.6%Run by Epoch AISimpleQA Verified	65.9%	Not yet published
Factuality. Epoch AI's run at max effort. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Answers in full. SpeechMap tested Kimi K3 through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. SpeechMap.AI
Kimi K2.7 Code	1%Run by Artificial AnalysisTerminal-Bench 4.0	47.8%	54%	87.9%	35%	24.5%	Not yet published	79.3%	Not yet published	36.5%Run by Epoch AISimpleQA Verified	Not yet published	63.1%Run by Artificial AnalysisIFBench
Agentic coding. Artificial Analysis ran Terminal-Bench 4.0 in its own agent harness at reasoning on effort. Artificial Analysis Agentic coding. Unusual. Artificial Analysis ran Terminal-Bench 4.0 in its own agent harness at reasoning on effort. Oct 31, 2026 Artificial Analysis Factuality. Epoch AI's run; Epoch prints no effort for it. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Instructions. Artificial Analysis ran IFBench directly at reasoning on effort. Artificial Analysis
Mistral Large 4	24.7%Run by Artificial AnalysisTerminal-Bench 4.0	Not yet published	Not yet published	Not yet published	Not yet published	57.8%	Not yet published	Not yet published	Not yet published	26%	Not yet published	Not yet published
Agentic coding. Artificial Analysis's run of the October 29 build at high effort in its own agent harness. Vals.ai has run only the October 6 preview build. Artificial Analysis Tool use and Factuality. Artificial Analysis's run of the October 29 build at high effort, its highest run of that build. Artificial Analysis
Mistral Medium 3.5	0%Run by Artificial AnalysisTerminal-Bench 4.0	40.2%	Not yet published	74.8%Run by Artificial AnalysisGPQA Diamond	13.8%	6.3%	Not yet published	69.3%	64.9%	23%	70.5%	68.8%Run by Artificial AnalysisIFBench
Agentic coding. Artificial Analysis's run at high effort in its own agent harness, from its Intelligence Index evaluations. It solved none of the tasks. Vals.ai has not run it. Artificial Analysis Agentic coding. Unusual. Artificial Analysis's run at high effort in its own agent harness, from its Intelligence Index evaluations. It solved none of the tasks. Vals.ai has not run it. Oct 28, 2026 Artificial Analysis Science. Artificial Analysis's run at high effort, the highest it ran for this release, from its Intelligence Index evaluations. Epoch AI has not run it. Artificial Analysis Answers in full. SpeechMap tested it through OpenRouter with reasoning on. With reasoning off it answers 87.9% in full. SpeechMap.AI Instructions. Artificial Analysis's IFBench run at high effort: the share of out-of-domain output constraints the model keeps. Scale AI has not run MultiChallenge on it. Artificial Analysis
Grok 4.7	28.8%	57.4%	53%	92.7%	43.1%	65.6%	Not yet published	76.7%	Not yet published	42.7%	64.1%	Not yet published
Answers in full. SpeechMap tested it through OpenRouter at medium reasoning and counted 34.4% of requests denied and 1.4% evaded. SpeechMap.AI
Grok 4.6	17.2%	56.5%	66%	93.2%	42.9%	66.7%	Not yet published	80.3%	Not yet published	48.9%Run by Epoch AISimpleQA Verified	83.6%	Not yet published
Factuality. Epoch AI's run at xhigh effort, the highest it ran; its high run scored 49.3. SimpleQA Verified prints no hallucination rate. Epoch AI Answers in full. SpeechMap tested Grok 4.6 through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. SpeechMap.AI
Grok 4.5	8.6%	55%	57.2%	93.4%	42.7%	57.9%	Not yet published	79.3%	80.4%	48.3%Run by Epoch AISimpleQA Verified	93.1%	Not yet published
Factuality. Epoch AI's run at high effort, the highest it ran. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Answers in full. SpeechMap tested Grok 4.5 through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. SpeechMap.AI
MiniMax M3.1 Flash	25.3%	53.4%	Not yet published	Not yet published	48%	72.5%	61.7%Reported by MiniMaxOSWorld 2.0	79.7%	79.1%	45.1%	Not yet published	Not yet published
Computer use. MiniMax's own figure on the August 2026 full task release, from its model card. MiniMax names neither the metric nor the harness nor the reasoning setting. MiniMax
Mercury 3	3%	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published	Not yet published
GLM-5.3-Flash	37.4%	51.6%	55.8%	90.2%	39.9%	60.4%	59.1%Reported by Z.aiOSWorld 2.0	80%	86%Run by Vals.aiMMMU-Pro	28.2%	65.4%	Not yet published
Computer use. Z.ai's own figure from its launch post. It does not name the OSWorld 2.0 task release, task set, harness or metric. No independent lab has run it. Z.ai Multimodal. Vals.ai's run, rank 23 on its MMMU Pro board, last updated September 1, 2026. Vals's variant offers 4 answer options, so it reads higher than the 10-option set. Vals.ai Answers in full. SpeechMap counted 28.9% of requests denied and 3.3% evaded. 2.4% failed at the provider. SpeechMap.AI
GLM-5.3	38.9%	59%	68.8%	90.9%	42.3%	62.2%	Not yet published	79.7%	Not yet published	32.3%	71.5%	Not yet published
Answers in full. SpeechMap counted 25.1% of requests denied and 3.3% evaded. SpeechMap.AI
GLM-5.2	1%Run by Artificial AnalysisTerminal-Bench 4.0	51.2%	59.2%	Not yet published	41.1%	28.4%	Not yet published	78.3%	Not yet published	34.2%Run by Epoch AISimpleQA Verified	Not yet published	73.3%Run by Artificial AnalysisIFBench
Agentic coding. Artificial Analysis ran Terminal-Bench 4.0 in its own agent harness at max effort. Artificial Analysis Agentic coding. Unusual. Artificial Analysis ran Terminal-Bench 4.0 in its own agent harness at max effort. Oct 31, 2026 Artificial Analysis Factuality. Epoch AI's run at max effort. SimpleQA Verified prints no hallucination rate, so none is shown beside it. Epoch AI Instructions. Artificial Analysis ran IFBench directly at max effort. Artificial Analysis
Muse Spark Open	6.6%	58.4%	69.8%	88.6%	43.9%	49.7%	80.4%Reported by MetaOSWorld-Verified	78.3%	80.9%	42%	Not yet published	74.2%
Agentic coding. Vals.ai ranked it 47th of 58 on its leaderboard of Oct 29, 2026. Artificial Analysis's own-harness run agrees, at 8.1. Vals.ai Agentic coding. Unusual. Vals.ai ranked it 47th of 58 on its leaderboard of Oct 29, 2026. Artificial Analysis's own-harness run agrees, at 8.1. Oct 29, 2026 Vals.ai Computer use. Meta's own figure from the Muse Spark Open model card, at xhigh effort. Meta does not state the harness or step budget. Meta Instructions. On Oct 29, 2026, Scale AI ranked it 6th, level with GPT-6.1 Terra. Scale AI SEAL
Muse Wonder	49.5%	60.3%	82.1%	Not yet published	50.1%	68.1%	72.6%Reported by Meta	80.7%	Not yet published	38%	42.7%	Not yet published
Agentic coding. Vals.ai ranked it 12th on its leaderboard of Oct 29, 2026, level with GPT-6.1 Terra at 49.49%. Vals.ai Computer use. Meta's own run at max effort in its own test framework on all 108 tasks: the average partial score. It fully completed 34.3% of tasks. Meta Answers in full. SpeechMap tested it through OpenRouter with reasoning on. The provider's API blocked 54.8% of requests; the model denied 1.4% and evaded 1.1%. SpeechMap.AI
Muse Spark 1.3	10.6%	58.8%	74%	93.5%Run by Artificial AnalysisGPQA Diamond	48.7%	57.9%	69.6%	83%	82%	38%	25.8%	Not yet published
Science. Artificial Analysis's run at max effort, the highest it ran for this release, from its Intelligence Index evaluations; its xhigh run scored 94.1. Epoch AI has not run it. Artificial Analysis Computer use. The OSWorld team's own run at max effort, with its batch tool and a 500-step budget. It fully completed 39.25% of tasks. OSWorld team Multimodal. Artificial Analysis ran MMMU-Pro on this release at xhigh effort only; its Max record carries no MMMU-Pro figure. Artificial Analysis Answers in full. SpeechMap tested Muse Spark 1.3 through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. SpeechMap.AI Answers in full. Unusual. SpeechMap tested Muse Spark 1.3 through its reasoning-on entry and reports complete, evasive, denial, and provider error shares. Oct 31, 2026 SpeechMap.AI
Llama 4 Scout	Not yet published	21.3%	7.8%Run by Epoch AIOTIS Mock AIME 2024-2025	51.8%	3.8%	0.2%	Not yet published	27.7%	52.9%	Not yet published	70.1%	39.5%Run by Artificial AnalysisIFBench
Math. Epoch AI's run of the instruction-tuned release. Llama 4 has no reasoning mode. Epoch AI Hard reasoning. Artificial Analysis publishes 3.8 for Llama 4 Scout, which has no reasoning mode, on this test. Artificial Analysis Tool use. Artificial Analysis publishes 0.2 for Llama 4 Scout on this test, far below its other results. Artificial Analysis Instructions. Artificial Analysis ran Llama 4 Scout, which has no reasoning mode, in its own evaluation harness. Artificial Analysis
Llama 4 Maverick	Not yet published	31.7%	20.6%Run by Epoch AIOTIS Mock AIME 2024-2025	67%	4.9%	0.2%	Not yet published	50%	62.1%	Not yet published	58.3%	43%Run by Artificial AnalysisIFBench
Math. Epoch AI's run of the FP8 instruction-tuned release. Llama 4 has no reasoning mode. Epoch AI Science. Epoch AI ran Meta's FP8 release of the weights. Epoch AI Hard reasoning. Artificial Analysis publishes 4.9 for Llama 4 Maverick, which has no reasoning mode, on this test. Artificial Analysis Hard reasoning. Unusual. Artificial Analysis publishes 4.9 for Llama 4 Maverick, which has no reasoning mode, on this test. Oct 28, 2026 Artificial Analysis Tool use. Artificial Analysis publishes 0.2 for Llama 4 Maverick on this test, far below its other results. Artificial Analysis Tool use. Unusual. Artificial Analysis publishes 0.2 for Llama 4 Maverick on this test, far below its other results. Oct 28, 2026 Artificial Analysis Instructions. Artificial Analysis ran Llama 4 Maverick, which has no reasoning mode, in its own evaluation harness. Artificial Analysis
Qwen3.8 27B	5.6%Run by Artificial AnalysisTerminal-Bench 4.0	46.6%	Not yet published	90.5%Run by Artificial AnalysisGPQA Diamond	33.9%	48.2%	84.3%Reported by AlibabaOSWorld-Verified	82%	76.3%	17.3%	Not yet published	79.5%Reported by AlibabaIFBench
Agentic coding. Artificial Analysis's run at xhigh effort in its own agent harness, from its Intelligence Index evaluations. Vals.ai has not run it. Artificial Analysis Agentic coding. Unusual. Artificial Analysis's run at xhigh effort in its own agent harness, from its Intelligence Index evaluations. Vals.ai has not run it. Nov 2, 2026 Artificial Analysis Science. Artificial Analysis's run at xhigh effort, the highest it ran for this release, from its Intelligence Index evaluations. Epoch AI has not run it. Artificial Analysis Computer use. Alibaba's own figure from the Qwen3.8 27B model card. It does not state the harness, step budget or reasoning effort. Alibaba Instructions. Alibaba's own figure from the Qwen3.8 27B model card. It does not state the reasoning effort. Scale AI has not run MultiChallenge on it. Alibaba
Qwen3.8 Max	38.9%	52.1%	65.6%	92.3%	43.1%	56.2%	Not yet published	80.3%	82.8%	29.5%	65.3%	Not yet published
Agentic coding. Vals.ai ranked the September 2 build 16th on its leaderboard of Oct 29, 2026. Vals.ai Answers in full. SpeechMap tested the September 2 build with reasoning on and counted 29.8% of requests denied and 2.1% evaded. SpeechMap.AI
MiMo-V2.6-Pro	31.3%	60.9%	Not yet published	Not yet published	49.4%	58.6%	82%Reported by XiaomiOSWorld-Verified	86.3%	Not yet published	34.8%	66.5%	Not yet published
Computer use. Xiaomi's own figure from the MiMo-V2.6-Pro model card. It does not state the harness, step budget or reasoning effort. No independent lab has run it. Xiaomi Answers in full. SpeechMap counted 29.7% of requests denied and 3.8% evaded. SpeechMap.AI
MiMo-V2.6-Flash	24.2%	51.3%	Not yet published	Not yet published	35.1%	64.1%	80.8%Reported by XiaomiOSWorld-Verified	74.3%	73.1%	27.2%	66.9%	Not yet published
Computer use. Xiaomi's own figure from the MiMo-V2.6-Flash model card. It does not state the harness, step budget or reasoning effort. No independent lab has run it. Xiaomi Answers in full. SpeechMap counted 30.1% of requests denied and 3.0% evaded. SpeechMap.AI
Ling 3.1 Flash	32.8%Run by Artificial AnalysisTerminal-Bench 4.0	56.8%	Not yet published	Not yet published	43.5%	67.5%	Not yet published	82.3%	Not yet published	26.5%	Not yet published	Not yet published
Agentic coding. Artificial Analysis's run with reasoning on in its own agent harness, from its Intelligence Index evaluations. Vals.ai has not run it. Artificial Analysis
Step 5 Preview	31.8%	58.9%	Not yet published	93.5%Reported by StepFunGPQA Diamond	46.5%	51%	Not yet published	88.3%	76.4%	36.4%	59.7%	Not yet published
Agentic coding. Vals.ai ranked it 22nd of 58 on its leaderboard of Oct 29, 2026. Vals.ai Science. StepFun's own figure from its launch table, at High reasoning, its highest setting. Epoch AI has not run it. StepFun Answers in full. SpeechMap ran it at medium reasoning through OpenRouter: 34.5% denied, 1.0% evaded, and 4.8% stopped by the provider's content filter. SpeechMap.AI

Axis

Agentic codingTerminal-Bench 4.0

Claude Opus 5.565.2%
Claude Sonnet 5.564.1%
Kimi K3.1 Pro63.6%
GPT-6 Astra59.6%
Claude Fable 5.158.1%
Gemini 4 Argon57.6%
GPT-6.1 Sol55.1%
Claude Opus 553.5%
GPT-6.1 Terra49.5%
Muse Wonder49.5%
GPT-6 Sol44.4%
Claude Fable 541.4%
GLM-5.338.9%
Qwen3.8 Max38.9%
GPT-5.6 Sol37.9%
GLM-5.3-Flash37.4%
Claude Haiku 5.536.4%
DeepSeek V4.1 Flash31.8%
Step 5 Preview31.8%
MiMo-V2.6-Pro31.3%
Grok 4.728.8%
MiniMax M3.1 Flash25.3%
MiMo-V2.6-Flash24.2%
Claude Opus 4.823.2%
GPT-5.6 Terra22.7%
Gemini 3.8 Flash19.7%
Kimi K317.2%
Grok 4.617.2%
DeepSeek V4 Pro14.1%
GPT-6 Luna13.6%
GPT-5.6 Luna11.6%
Claude Sonnet 510.6%
Muse Spark 1.310.6%
Grok 4.58.6%
Muse Spark Open6.6%
Mercury 33%
Gemini 3.1 Pro Preview2.5%
Ling 3.1 Flash32.8%Artificial Analysis's run, Terminal-Bench 4.0
Mistral Large 424.7%Artificial Analysis's run, Terminal-Bench 4.0
GPT-5.514.6%Artificial Analysis's run, Terminal-Bench 4.0
gpt-oss-120b14.2%Terminal-Bench's run, Terminal-Bench 2.0
Gemini 3.6 Flash7.1%Artificial Analysis's run, Terminal-Bench 4.0
Qwen3.8 27B5.6%Artificial Analysis's run, Terminal-Bench 4.0
Claude Sonnet 4.63%Artificial Analysis's run, Terminal-Bench 4.0
Kimi K2.7 Code1%Artificial Analysis's run, Terminal-Bench 4.0
GLM-5.21%Artificial Analysis's run, Terminal-Bench 4.0
Mistral Medium 3.50%Artificial Analysis's run, Terminal-Bench 4.0

The other models in the grid have no result on Terminal-Bench 4.0 yet.

CodeSciCode

The other models in the grid have no result on SciCode yet.

MathFrontierMath Tiers 1-3

GPT-6.1 Sol93.7%
GPT-6 Astra93.7%
Claude Opus 5.591.2%
GPT-6.1 Terra90.9%
Claude Fable 5.190.2%
GPT-6 Sol89.8%
GPT-5.6 Sol89.1%
Claude Sonnet 5.588.8%
Claude Fable 587%
GPT-5.6 Terra86%
Gemini 4 Argon86%
Claude Opus 585.6%
GPT-5.585.3%
GPT-5.6 Luna82.1%
Muse Wonder82.1%
Claude Opus 4.880%
GPT-6 Luna78.9%
GPT-5.478.6%
Claude Haiku 5.575.1%
Muse Spark 1.374%
Kimi K372.2%
DeepSeek V4.1 Flash71.2%
Claude Opus 4.770.2%
Muse Spark Open69.8%
GLM-5.368.8%
Gemini 3.8 Flash68.4%
Claude Opus 4.666%
Grok 4.666%
Claude Sonnet 565.6%
Qwen3.8 Max65.6%
DeepSeek V4 Pro64.6%
Gemini 3.1 Pro Preview59.6%
GLM-5.259.2%
Gemini 3.6 Flash58.9%
Grok 4.557.2%
GLM-5.3-Flash55.8%
Kimi K2.7 Code54%
Grok 4.753%
Claude Opus 4.534.4%
Claude Sonnet 4.523.9%
gpt-oss-120b88.9%Epoch AI's run, OTIS Mock AIME 2024-2025
Claude Sonnet 4.671.1%Epoch AI's run, OTIS Mock AIME 2024-2025
Claude Haiku 4.566.7%Epoch AI's run, OTIS Mock AIME 2024-2025
Llama 4 Maverick20.6%Epoch AI's run, OTIS Mock AIME 2024-2025
Llama 4 Scout7.8%Epoch AI's run, OTIS Mock AIME 2024-2025

The other models in the grid have no result on FrontierMath Tiers 1-3 yet.

ScienceGPQA Diamond

Gemini 4 Argon96.4%
GPT-6 Astra95.8%
Claude Sonnet 5.595.6%
GPT-6.1 Sol95.4%
Gemini 3.8 Flash95.4%
Gemini 3.1 Pro Preview94.4%
GPT-6.1 Terra94.3%
GPT-6 Sol94.3%
Gemini 3.6 Flash94.1%
Claude Opus 593.9%
GPT-5.6 Sol93.5%
Grok 4.593.4%
GPT-5.6 Terra93.3%
GPT-5.493.3%
DeepSeek V4.1 Flash93.2%
Grok 4.693.2%
Kimi K393.1%
Grok 4.792.7%
Qwen3.8 Max92.3%
DeepSeek V4 Pro91.7%
GPT-5.6 Luna91.6%
Claude Opus 4.891%
GLM-5.390.9%
Claude Opus 5.590.6%
GPT-6 Luna90.5%
GLM-5.3-Flash90.2%
Claude Haiku 5.589.6%
Muse Spark Open88.6%
Claude Opus 4.688.4%
Kimi K2.7 Code87.9%
Claude Opus 4.786.4%
Claude Opus 4.586%
Claude Fable 585.9%
Claude Sonnet 4.582.3%
Claude Sonnet 580.3%
Claude Sonnet 4.678.8%
gpt-oss-120b75.8%
Claude Haiku 4.571.2%
Llama 4 Maverick67%
Llama 4 Scout51.8%
Kimi K3.1 Pro94.2%Moonshot AI's own figure
GPT-5.593.5%Artificial Analysis's run
Muse Spark 1.393.5%Artificial Analysis's run, GPQA Diamond
Step 5 Preview93.5%StepFun's own figure, GPQA Diamond
Claude Fable 5.193.4%Vals.ai's run
GPT-5.3-Codex91.5%Artificial Analysis's run
Qwen3.8 27B90.5%Artificial Analysis's run, GPQA Diamond
Gemma 4 31B85.7%Artificial Analysis's run, GPQA Diamond
Mistral Medium 3.574.8%Artificial Analysis's run, GPQA Diamond

The other models in the grid have no result on GPQA Diamond yet.

Hard reasoningHumanity's Last Exam

The other models in the grid have no result on Humanity's Last Exam yet.

Tool useAutomationBench-AA

The other models in the grid have no result on AutomationBench-AA yet.

Computer useOSWorld 2.1

Claude Opus 577.7%
Muse Spark 1.369.6%
Qwen3.8 27B84.3%Alibaba's own figure, OSWorld-Verified
Gemini 3.6 Flash83%Google's own figure, OSWorld-Verified
MiMo-V2.6-Pro82%Xiaomi's own figure, OSWorld-Verified
Claude Opus 5.581.8%Anthropic's own figure
MiMo-V2.6-Flash80.8%Xiaomi's own figure, OSWorld-Verified
Claude Fable 5.180.7%Anthropic's own figure
Muse Spark Open80.4%Meta's own figure, OSWorld-Verified
Claude Sonnet 5.580.1%Anthropic's own figure
Kimi K3.1 Pro79.6%Moonshot AI's own figure, OSWorld 2.1
GPT-6.1 Sol78%OpenAI's own figure, OSWorld 2.1
GPT-6 Astra75.8%Meta's run
GPT-6.1 Terra74.2%OpenAI's own figure, OSWorld 2.1
Muse Wonder72.6%Meta's own figure
Claude Haiku 5.572.4%Anthropic's own figure, OSWorld 2.1
DeepSeek V4.1 Flash69.3%Moonshot AI's run, OSWorld 2.1
Gemini 4 Argon69.2%Google's own figure, OSWorld 2.0
GPT-6 Sol64.4%OpenAI's own figure, OSWorld 2.0
GPT-5.6 Sol62.7%OSWorld team's run, OSWorld 2.0
MiniMax M3.1 Flash61.7%MiniMax's own figure, OSWorld 2.0
GLM-5.3-Flash59.1%Z.ai's own figure, OSWorld 2.0
Gemini 3.8 Flash59%Google's own figure, OSWorld 2.0
Claude Sonnet 557%Anthropic's own figure
Claude Opus 4.854.8%OSWorld team's run, OSWorld 2.0
GPT-5.549.5%OSWorld team's run, OSWorld 2.0
GPT-6 Luna48.9%Anthropic's run, OSWorld 2.1
Claude Opus 4.748.9%OSWorld team's run, OSWorld 2.0
Claude Sonnet 4.641.5%OSWorld team's run, OSWorld 2.0

The other models in the grid have no result on OSWorld 2.1 yet.

Long contextAA-LCR v1.1

The other models in the grid have no result on AA-LCR v1.1 yet.

MultimodalMMMU-Pro

The other models in the grid have no result on MMMU-Pro yet.

FactualityAA-Omniscience Accuracy

Claude Fable 5.168.4%
Claude Opus 5.565.5%
Claude Fable 565.4%
GPT-6.1 Sol61.8%
GPT-6 Astra61.6%
Claude Opus 560.9%
Gemini 3.1 Pro Preview54.9%
GPT-6 Sol54.5%
Claude Sonnet 5.551.5%
Gemini 3.8 Flash50.5%
Claude Opus 4.748.9%
Kimi K3.1 Pro48.9%
Claude Opus 4.848.8%
Gemini 4 Argon48.3%
GPT-6.1 Terra47.6%
Claude Opus 4.647%
Claude Opus 4.546.6%
DeepSeek V4 Pro45.2%
MiniMax M3.1 Flash45.1%
Grok 4.742.7%
GPT-5.6 Terra42.3%
Muse Spark Open42%
Claude Sonnet 4.640.9%
Claude Sonnet 540.1%
GPT-6 Luna39.6%
DeepSeek V4.1 Flash39.6%
Muse Wonder38%
Muse Spark 1.338%
Step 5 Preview36.4%
MiMo-V2.6-Pro34.8%
Claude Haiku 5.534.7%
Claude Sonnet 4.532.9%
GLM-5.332.3%
Qwen3.8 Max29.5%
GLM-5.3-Flash28.2%
MiMo-V2.6-Flash27.2%
Ling 3.1 Flash26.5%
Mistral Large 426%
Mistral Medium 3.523%
Claude Haiku 4.518%
Qwen3.8 27B17.3%
GPT-5.6 Sol69.7%Epoch AI's run, SimpleQA Verified
Gemini 3.6 Flash66.2%Epoch AI's run, SimpleQA Verified
GPT-5.563%Epoch AI's run, SimpleQA Verified
Kimi K350.6%Epoch AI's run, SimpleQA Verified
Grok 4.648.9%Epoch AI's run, SimpleQA Verified
Grok 4.548.3%Epoch AI's run, SimpleQA Verified
GPT-5.6 Luna41%Epoch AI's run, SimpleQA Verified
Kimi K2.7 Code36.5%Epoch AI's run, SimpleQA Verified
GLM-5.234.2%Epoch AI's run, SimpleQA Verified
gpt-oss-120b16.8%OpenAI's own figure, SimpleQA
Gemma 4 31B10.4%Epoch AI's run, SimpleQA Verified

The other models in the grid have no result on AA-Omniscience Accuracy yet.

Instruction followingMultiChallenge

GPT-6.1 Sol82%
Gemini 4 Argon78.7%
DeepSeek V4.1 Flash74.9%
GPT-6.1 Terra74.2%
Muse Spark Open74.2%
Gemini 3.1 Pro Preview71.4%
Claude Haiku 5.567.7%
Claude Opus 4.559%
Claude Sonnet 4.555.3%
Claude Haiku 4.550.5%
gpt-oss-120b45.3%
Claude Opus 4.637.2%
Qwen3.8 27B79.5%Alibaba's own figure, IFBench
GPT-5.575.9%Artificial Analysis's run, IFBench
Gemma 4 31B75.6%Artificial Analysis's run, IFBench
GPT-5.3-Codex75.4%Artificial Analysis's run, IFBench
GPT-5.473.9%Artificial Analysis's run, IFBench
GLM-5.273.3%Artificial Analysis's run, IFBench
GPT-5.6 Sol72.7%Artificial Analysis's run, IFBench
GPT-5.6 Terra71.2%Artificial Analysis's run, IFBench
Mistral Medium 3.568.8%Artificial Analysis's run, IFBench
Claude Fable 563.5%Artificial Analysis's run, IFBench
Kimi K2.7 Code63.1%Artificial Analysis's run, IFBench
Claude Opus 4.862.2%Artificial Analysis's run, IFBench
Claude Opus 4.758.6%Artificial Analysis's run, IFBench
Claude Sonnet 4.656.6%Artificial Analysis's run, IFBench
Llama 4 Maverick43%Artificial Analysis's run, IFBench
Llama 4 Scout39.5%Artificial Analysis's run, IFBench

The other models in the grid have no result on MultiChallenge yet.

Releases

Every GPT Terra release we track, newest first.

What practitioners report

Strongest on Speed, weakest on Quality and reasoning.

Each figure is the share of accounts naming GPT-6.1 Terra that speak well of it on that axis. An axis is scored once praise and concerns are both on record from at least five accounts by three authors. An account from the maker, a company that sells or serves the model, or the maker's expert program is listed and counts toward no figure.

Share of practitioner accounts favorable to GPT-6.1 Terra on each of seven axes, with the accounts behind each figure
Axis	Favorable	Accounts
Quality and reasoningWhether its answers and code are right on the author's own task: reasoning, accuracy, instruction following, and consistency.	33%	3 favorable, 6 not, 8 authors
Tool use and structured outputWhether it calls tools correctly, returns valid structured output, and keeps agent loops on track.	40%	2 favorable, 3 not, 5 authors
Long contextHow it holds up over long documents, large codebases, long sessions, and many images or PDFs.	Only concerns on record	0 favorable, 1 not, 1 author
Developer experienceHow it is to build with: prompting, steering, effort settings, and support across SDKs and providers.	Only concerns on record	0 favorable, 2 not, 2 authors
CostWhat the work costs: price per task, token use, and what caching and batch save.	80%	8 favorable, 2 not, 9 authors
SpeedResponse time, time to first token, and throughput.	83%	5 favorable, 1 not, 5 authors
ReliabilityWhether requests keep succeeding: overloaded errors, outages, rate limits, and changes that broke what worked.	Only concerns on record	0 favorable, 1 not, 1 author

Every model on these seven axes

Share of accounts favorable on each axis for GPT-6.1 Terra and every other model GILD has accounts for
Axis	GPT-6.1 Terra	GPT-6.1 Sol	GPT-6 Luna	GPT-6 Astra	GPT-5.6 Sol	GPT-5.6 Terra	GPT-5.6 Luna	GPT-5.5	GPT-5.4	gpt-oss-120b	GPT-6 Sol	GPT-5.3-Codex	Claude Haiku 5.5	Claude Sonnet 5.5	Claude Opus 5.5	Claude Fable 5.1	Claude Opus 5	Claude Sonnet 5	Claude Fable 5	Claude Opus 4.8	Claude Opus 4.7	Claude Sonnet 4.6	Claude Opus 4.6	Claude Opus 4.5	Claude Haiku 4.5	Claude Sonnet 4.5	Gemini 4 Argon	Gemini 3.8 Flash	Gemma 4 31B	Gemini 3.1 Pro Preview	Gemini 3.6 Flash	DeepSeek V4.1 Flash	DeepSeek V4 Pro	Kimi K3.1 Pro	Kimi K3	Kimi K2.7 Code	Mistral Large 4	Mistral Medium 3.5	Grok 4.7	Grok 4.6	Grok 4.5	MiniMax M3.1 Flash	Mercury 3	GLM-5.3-Flash	GLM-5.3	GLM-5.2	Muse Spark Open	Muse Wonder	Muse Spark 1.3	Llama 4 Scout	Llama 4 Maverick	Qwen3.8 27B	Qwen3.8 Max	MiMo-V2.6-Pro	MiMo-V2.6-Flash	Ling 3.1 Flash	Step 5 Preview
Quality and reasoning	33%3 favorable, 6 not, 8 authors	55%12 favorable, 10 not, 22 authors	55%16 favorable, 13 not, 29 authors	44%12 favorable, 15 not, 26 authors	33%3 favorable, 6 not, 9 authors	38%3 favorable, 5 not, 8 authors	Needs 5 accounts from 3 authors3 favorable, 1 not, 4 authors	33%4 favorable, 8 not, 12 authors	33%2 favorable, 4 not, 6 authors	58%22 favorable, 16 not, 38 authors	31%10 favorable, 22 not, 32 authors	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	73%19 favorable, 7 not, 26 authors	64%14 favorable, 8 not, 22 authors	68%17 favorable, 8 not, 25 authors	54%15 favorable, 13 not, 28 authors	50%6 favorable, 6 not, 12 authors	56%9 favorable, 7 not, 16 authors	68%13 favorable, 6 not, 18 authors	68%17 favorable, 8 not, 25 authors	57%12 favorable, 9 not, 21 authors	67%12 favorable, 6 not, 18 authors	70%19 favorable, 8 not, 27 authors	62%13 favorable, 8 not, 20 authors	58%14 favorable, 10 not, 24 authors	69%11 favorable, 5 not, 16 authors	63%22 favorable, 13 not, 35 authors	56%27 favorable, 21 not, 45 authors	80%24 favorable, 6 not, 28 authors	56%20 favorable, 16 not, 36 authors	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	61%25 favorable, 16 not, 40 authors	Only concerns on record0 favorable, 4 not, 4 authors	59%20 favorable, 14 not, 32 authors	Only praise on record3 favorable, 0 not, 3 authors	50%3 favorable, 3 not, 6 authors	58%11 favorable, 8 not, 18 authors	36%5 favorable, 9 not, 14 authors	43%16 favorable, 21 not, 36 authors	86%6 favorable, 1 not, 7 authors	60%3 favorable, 2 not, 5 authors	62%13 favorable, 8 not, 21 authors	86%18 favorable, 3 not, 21 authors	67%14 favorable, 7 not, 21 authors	76%47 favorable, 15 not, 61 authors	71%5 favorable, 2 not, 7 authors	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	20%1 favorable, 4 not, 5 authors	Only praise on record2 favorable, 0 not, 2 authors	53%8 favorable, 7 not, 15 authors	44%8 favorable, 10 not, 17 authors	71%24 favorable, 10 not, 34 authors	57%4 favorable, 3 not, 7 authors	22%2 favorable, 7 not, 9 authors	40%4 favorable, 6 not, 10 authors	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors
Tool use and structured output	40%2 favorable, 3 not, 5 authors	40%4 favorable, 6 not, 10 authors	33%3 favorable, 6 not, 9 authors	62%8 favorable, 5 not, 13 authors	No accounts yet	Only concerns on record0 favorable, 2 not, 2 authors	Only praise on record1 favorable, 0 not, 1 author	Only praise on record1 favorable, 0 not, 1 author	No accounts yet	27%4 favorable, 11 not, 15 authors	56%5 favorable, 4 not, 9 authors	No accounts yet	67%6 favorable, 3 not, 9 authors	40%4 favorable, 6 not, 10 authors	46%6 favorable, 7 not, 13 authors	30%3 favorable, 7 not, 10 authors	Needs 5 accounts from 3 authors3 favorable, 1 not, 4 authors	70%7 favorable, 3 not, 10 authors	57%4 favorable, 3 not, 7 authors	50%4 favorable, 4 not, 8 authors	50%4 favorable, 4 not, 8 authors	29%2 favorable, 5 not, 7 authors	33%4 favorable, 8 not, 12 authors	70%7 favorable, 3 not, 10 authors	43%3 favorable, 4 not, 6 authors	50%3 favorable, 3 not, 6 authors	25%2 favorable, 6 not, 8 authors	47%7 favorable, 8 not, 15 authors	67%4 favorable, 2 not, 6 authors	54%7 favorable, 6 not, 12 authors	Only concerns on record0 favorable, 1 not, 1 author	59%10 favorable, 7 not, 17 authors	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors3 favorable, 1 not, 4 authors	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	38%5 favorable, 8 not, 13 authors	Only concerns on record0 favorable, 5 not, 5 authors	38%6 favorable, 10 not, 16 authors	Only praise on record2 favorable, 0 not, 2 authors	Only concerns on record0 favorable, 1 not, 1 author	67%4 favorable, 2 not, 6 authors	80%8 favorable, 2 not, 10 authors	33%2 favorable, 4 not, 6 authors	57%4 favorable, 3 not, 7 authors	No accounts yet	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	Needs 5 accounts from 3 authors1 favorable, 2 not, 3 authors	Only concerns on record0 favorable, 4 not, 4 authors	46%6 favorable, 7 not, 13 authors	Only praise on record1 favorable, 0 not, 1 author	Only concerns on record0 favorable, 5 not, 5 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author
Long context	Only concerns on record0 favorable, 1 not, 1 author	38%3 favorable, 5 not, 8 authors	20%1 favorable, 4 not, 5 authors	27%3 favorable, 8 not, 11 authors	No accounts yet	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	25%2 favorable, 6 not, 8 authors	Only concerns on record0 favorable, 3 not, 3 authors	No accounts yet	29%2 favorable, 5 not, 7 authors	60%3 favorable, 2 not, 5 authors	80%8 favorable, 2 not, 10 authors	63%5 favorable, 3 not, 8 authors	67%4 favorable, 2 not, 6 authors	50%3 favorable, 3 not, 6 authors	33%2 favorable, 4 not, 6 authors	60%6 favorable, 4 not, 10 authors	50%4 favorable, 4 not, 7 authors	71%5 favorable, 2 not, 7 authors	64%7 favorable, 4 not, 11 authors	40%4 favorable, 6 not, 10 authors	40%2 favorable, 3 not, 5 authors	50%4 favorable, 4 not, 8 authors	25%2 favorable, 6 not, 8 authors	38%5 favorable, 8 not, 13 authors	20%1 favorable, 4 not, 5 authors	44%7 favorable, 9 not, 16 authors	Only praise on record1 favorable, 0 not, 1 author	58%7 favorable, 5 not, 12 authors	No accounts yet	Needs 5 accounts from 3 authors3 favorable, 1 not, 4 authors	Only praise on record1 favorable, 0 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	No accounts yet	42%5 favorable, 7 not, 12 authors	No accounts yet	No accounts yet	40%2 favorable, 3 not, 5 authors	60%3 favorable, 2 not, 5 authors	No accounts yet	40%2 favorable, 3 not, 5 authors	No accounts yet	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 3 not, 3 authors	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	75%6 favorable, 2 not, 8 authors	Only praise on record1 favorable, 0 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	No accounts yet	No accounts yet
Developer experience	Only concerns on record0 favorable, 2 not, 2 authors	27%3 favorable, 8 not, 11 authors	25%2 favorable, 6 not, 8 authors	41%7 favorable, 10 not, 17 authors	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	No accounts yet	No accounts yet	14%1 favorable, 6 not, 7 authors	25%3 favorable, 9 not, 12 authors	No accounts yet	25%2 favorable, 6 not, 8 authors	38%5 favorable, 8 not, 13 authors	38%6 favorable, 10 not, 16 authors	50%7 favorable, 7 not, 14 authors	25%3 favorable, 9 not, 12 authors	29%2 favorable, 5 not, 7 authors	25%3 favorable, 9 not, 12 authors	53%10 favorable, 9 not, 19 authors	29%4 favorable, 10 not, 14 authors	57%4 favorable, 3 not, 7 authors	45%5 favorable, 6 not, 11 authors	67%8 favorable, 4 not, 12 authors	Needs 5 accounts from 3 authors2 favorable, 2 not, 4 authors	50%3 favorable, 3 not, 6 authors	33%3 favorable, 6 not, 9 authors	39%7 favorable, 11 not, 17 authors	Only concerns on record0 favorable, 7 not, 7 authors	40%8 favorable, 12 not, 20 authors	No accounts yet	35%8 favorable, 15 not, 23 authors	No accounts yet	9%1 favorable, 10 not, 11 authors	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	13%2 favorable, 14 not, 16 authors	Only concerns on record0 favorable, 9 not, 9 authors	25%5 favorable, 15 not, 20 authors	No accounts yet	No accounts yet	Only concerns on record0 favorable, 7 not, 7 authors	11%1 favorable, 8 not, 9 authors	Only concerns on record0 favorable, 3 not, 3 authors	33%5 favorable, 10 not, 15 authors	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	Only concerns on record0 favorable, 6 not, 6 authors	Only concerns on record0 favorable, 1 not, 1 author	17%1 favorable, 5 not, 6 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	No accounts yet
Cost	80%8 favorable, 2 not, 9 authors	74%14 favorable, 5 not, 19 authors	79%22 favorable, 6 not, 26 authors	29%7 favorable, 17 not, 23 authors	57%4 favorable, 3 not, 7 authors	Only concerns on record0 favorable, 1 not, 1 author	50%3 favorable, 3 not, 6 authors	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	63%5 favorable, 3 not, 8 authors	60%3 favorable, 2 not, 5 authors	55%12 favorable, 10 not, 21 authors	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	69%20 favorable, 9 not, 29 authors	45%10 favorable, 12 not, 22 authors	56%15 favorable, 12 not, 27 authors	20%4 favorable, 16 not, 20 authors	43%3 favorable, 4 not, 7 authors	39%7 favorable, 11 not, 17 authors	8%1 favorable, 11 not, 12 authors	40%6 favorable, 9 not, 14 authors	50%7 favorable, 7 not, 13 authors	50%6 favorable, 6 not, 12 authors	33%7 favorable, 14 not, 20 authors	53%8 favorable, 7 not, 15 authors	29%4 favorable, 10 not, 12 authors	38%3 favorable, 5 not, 8 authors	35%7 favorable, 13 not, 20 authors	58%14 favorable, 10 not, 23 authors	Only praise on record7 favorable, 0 not, 7 authors	50%8 favorable, 8 not, 16 authors	Only concerns on record0 favorable, 1 not, 1 author	66%25 favorable, 13 not, 37 authors	Only praise on record2 favorable, 0 not, 2 authors	40%10 favorable, 15 not, 25 authors	71%5 favorable, 2 not, 7 authors	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	27%3 favorable, 8 not, 11 authors	Only concerns on record0 favorable, 8 not, 8 authors	19%5 favorable, 21 not, 26 authors	20%1 favorable, 4 not, 5 authors	Only praise on record4 favorable, 0 not, 4 authors	80%12 favorable, 3 not, 15 authors	50%6 favorable, 6 not, 12 authors	76%16 favorable, 5 not, 21 authors	35%8 favorable, 15 not, 23 authors	71%5 favorable, 2 not, 7 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 9 not, 9 authors	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	Only praise on record1 favorable, 0 not, 1 author	Only praise on record1 favorable, 0 not, 1 author	Needs 5 accounts from 3 authors2 favorable, 2 not, 4 authors	Only concerns on record0 favorable, 2 not, 2 authors	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	Only praise on record3 favorable, 0 not, 3 authors	Only praise on record1 favorable, 0 not, 1 author	Only praise on record1 favorable, 0 not, 1 author
Speed	83%5 favorable, 1 not, 5 authors	37%7 favorable, 12 not, 19 authors	45%5 favorable, 6 not, 11 authors	41%7 favorable, 10 not, 17 authors	Only praise on record1 favorable, 0 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	80%4 favorable, 1 not, 5 authors	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	No accounts yet	85%23 favorable, 4 not, 27 authors	57%4 favorable, 3 not, 7 authors	Only praise on record1 favorable, 0 not, 1 author	74%14 favorable, 5 not, 19 authors	88%14 favorable, 2 not, 16 authors	46%6 favorable, 7 not, 13 authors	44%4 favorable, 5 not, 9 authors	60%3 favorable, 2 not, 5 authors	44%4 favorable, 5 not, 9 authors	57%4 favorable, 3 not, 7 authors	31%4 favorable, 9 not, 13 authors	63%5 favorable, 3 not, 8 authors	17%1 favorable, 5 not, 6 authors	36%4 favorable, 7 not, 11 authors	73%8 favorable, 3 not, 11 authors	78%7 favorable, 2 not, 9 authors	75%6 favorable, 2 not, 8 authors	45%5 favorable, 6 not, 11 authors	61%19 favorable, 12 not, 31 authors	53%10 favorable, 9 not, 17 authors	35%6 favorable, 11 not, 16 authors	Needs 5 accounts from 3 authors2 favorable, 1 not, 3 authors	68%21 favorable, 10 not, 30 authors	Only concerns on record0 favorable, 1 not, 1 author	21%4 favorable, 15 not, 19 authors	Only concerns on record0 favorable, 2 not, 2 authors	No accounts yet	21%3 favorable, 11 not, 14 authors	Only concerns on record0 favorable, 2 not, 2 authors	29%5 favorable, 12 not, 17 authors	Needs 5 accounts from 3 authors1 favorable, 2 not, 3 authors	Only praise on record2 favorable, 0 not, 2 authors	44%4 favorable, 5 not, 9 authors	65%13 favorable, 7 not, 20 authors	33%3 favorable, 6 not, 9 authors	20%3 favorable, 12 not, 15 authors	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	Needs 5 accounts from 3 authors3 favorable, 1 not, 4 authors	No accounts yet	33%7 favorable, 14 not, 21 authors	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	Only concerns on record0 favorable, 4 not, 4 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 2 not, 2 authors	Only concerns on record0 favorable, 1 not, 1 author
Reliability	Only concerns on record0 favorable, 1 not, 1 author	13%1 favorable, 7 not, 8 authors	Only concerns on record0 favorable, 5 not, 5 authors	14%2 favorable, 12 not, 14 authors	Only concerns on record0 favorable, 4 not, 4 authors	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	Only concerns on record0 favorable, 2 not, 2 authors	Only concerns on record0 favorable, 4 not, 4 authors	Only concerns on record0 favorable, 4 not, 4 authors	17%1 favorable, 5 not, 6 authors	Only concerns on record0 favorable, 3 not, 3 authors	20%1 favorable, 4 not, 5 authors	18%2 favorable, 9 not, 11 authors	13%1 favorable, 7 not, 8 authors	9%1 favorable, 10 not, 11 authors	13%1 favorable, 7 not, 8 authors	63%5 favorable, 3 not, 7 authors	22%2 favorable, 7 not, 9 authors	17%2 favorable, 10 not, 12 authors	18%2 favorable, 9 not, 11 authors	17%1 favorable, 5 not, 6 authors	20%3 favorable, 12 not, 15 authors	38%6 favorable, 10 not, 16 authors	29%2 favorable, 5 not, 7 authors	40%4 favorable, 6 not, 10 authors	5%1 favorable, 18 not, 19 authors	15%3 favorable, 17 not, 20 authors	Only concerns on record0 favorable, 5 not, 5 authors	14%3 favorable, 19 not, 22 authors	Only concerns on record0 favorable, 1 not, 1 author	42%5 favorable, 7 not, 12 authors	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	8%1 favorable, 12 not, 13 authors	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 1 not, 2 authors	38%3 favorable, 5 not, 8 authors	Only concerns on record0 favorable, 4 not, 4 authors	22%2 favorable, 7 not, 9 authors	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	60%3 favorable, 2 not, 5 authors	33%2 favorable, 4 not, 5 authors	Only concerns on record0 favorable, 7 not, 7 authors	Only concerns on record0 favorable, 4 not, 4 authors	Only concerns on record0 favorable, 1 not, 1 author	No accounts yet	Only concerns on record0 favorable, 2 not, 2 authors	Only concerns on record0 favorable, 1 not, 1 author	Needs 5 accounts from 3 authors1 favorable, 3 not, 4 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 9 not, 9 authors	Only concerns on record0 favorable, 1 not, 1 author	Only concerns on record0 favorable, 2 not, 2 authors	No accounts yet	No accounts yet	Only concerns on record0 favorable, 1 not, 1 author

The 21 accounts behind GPT-6.1 Terra's figures

GILD searched Hacker News, GitHub and X for accounts that name GPT-6.1 Terra.

Quality and reasoning 33%

Favorable
GPT-6.1 Terra passed 40 of 43 questions (93%) for $1.31 for the whole exam, half the cost of GPT-6.1 Sol, which scored 43.
GPT-6.1 Terra Benchmark Results for Data Analysis. Plotly, maker of the Plotly Studio analytics agent and the Dash framework, self-described, Oct 26, 2026. GPT-6.1 Terra through OpenRouter at its default reasoning setting on Plotly's private 43-question data analytics exam.
Favorable
GPT-6.1 Terra scored 84, one point over Grok 4.7 (83)
GPT-6.1 Terra. thefourthchime, Hacker News commenter, self-described, Oct 27, 2026. GPT-6.1 Terra on PacBench, a Pac-Man build bake-off, against Grok 4.7.
Favorable
Same car. Better result from Terra.
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against GPT-6.1 Sol, both at Extra High, on the same car computer use task.
Unfavorable
But Haiku produced the more realistic result — better lighting, fog and fire.
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against Claude Haiku 5.5 on the same build.
Unfavorable
less keen on refactoring when the grid scales to 22×22.
X post by @RichDevLab. Richard Thomas (RichDevLab), architect and author of farmerbench.dev per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in Codex at medium effort on three 6-hour FarmerBench seeds.
Unfavorable
blind test vs the model we already run: - lost 14 to 3
X post by @RimasXYZ. Rimas (RimasXYZ), building adrio.ai per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in a blind test on the author's ad prompt agent against the model already in production.
Unfavorable
Tried 6.1 Terra in codex, not very impressed with it, I 'll stick to Sol.
X post by @Bi11Leou. Bill Leoutsakos (Bi11Leou), computer engineering at Cambridge per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra in Codex.
Unfavorable
It still bypass the "do not delete anything without my tell you so" rule.
X post by @xVelns. xVelns, X user, self-described, Oct 27, 2026. GPT-6.1 Terra at Max effort with a do-not-delete rule.
Unfavorable
even this watered down version from @opencode is faster, cheaper and more capable than @OpenAI GPT 6.1-Terra High
X post by @GCC_Trader. Camel Trader (GCC_Trader), X user, self-described, Oct 27, 2026. GPT-6.1 Terra (GPT 6.1-Terra) at High effort against DeepSeek in OpenCode.

Tool use and structured output 40%

Favorable
The new Terra 6.1 is pretty good at doing user testing as well via computer use, going through all the features and noting any bugs or inconsistencies.
Genie 4. pigpop, Hacker News commenter, self-described, Oct 28, 2026. GPT-6.1 Terra (Terra 6.1) user testing a game project through computer use.
Unfavorable
I ask it to implement something, but it doesn't follow through. It keeps stopping without taking the next concrete step.
gpt 6.1 terra bad. hybridapp323, GitHub issue author, self-described, Oct 28, 2026. GPT-6.1 Terra (gpt 6.1 terra) in Codex CLI v0.165 implementing features.
Unfavorable
I asked GPT 6.1 Terra to perform a test run on a personal Computer Use benchmark, but it stopped very early because it was "too complex". This has happened 3 times now.
X post by @ishuagra02. Ishu Agrawal (ishuagra02), AI news and demos poster per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra on the author's personal computer use benchmark.
Favorable
when a locked hat crashed its script on seed 3, it read the docs, wrote a one-line fix and recovered in a minute.
X post by @RichDevLab. Richard Thomas (RichDevLab), architect and author of farmerbench.dev per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in Codex at medium effort on three 6-hour FarmerBench seeds.
Unfavorable
Stops tasks randomly. Was instructed to launch subagents but claimed GPT Terra 6.1 isn't an available model so it stopped and gave up.
X post by @synthshareai. Petra (synthshareai), GenAI explorer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra (GPT Terra 6.1) in Codex launching subagents.

Long context Only concerns on record

Unfavorable
GPT Terra 6.1 worked great for a hour or two for me. Then it turned to slop.
X post by @GDWizard00. GDWizard (GDWizard00), AI, Web3 and gaming builder per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra (GPT Terra 6.1) over a session of one repeated task.

Developer experience Only concerns on record

Unfavorable
This appears to be a stale or incorrect model label in the VS Code extension during the rollout
VS Code Codex showed GPT-6 Terra while session metadata used gpt-6.1-terra during Oct 26 rollout. pearson-bailey, GitHub issue author, self-described, Oct 26, 2026. GPT-6.1 Terra (gpt-6.1-terra) in the Codex VS Code extension at xhigh effort during the October 26 rollout.
Unfavorable
Hermes saves `agent.reasoning_effort: max` and builds `reasoning.effort: xhigh` for it.
[Bug]: GPT-6.1 Terra max reasoning is silently clamped to xhigh on Codex Responses. Tavitsri, GitHub issue author, self-described, Oct 27, 2026. GPT-6.1 Terra (gpt-6.1-terra) through Hermes Agent's Codex Responses route.

Cost 80%

Favorable
It also joins the cost-accuracy frontier: no model we've tested is both cheaper and more accurate.
GPT-6.1 Terra Benchmark Results for Data Analysis. Plotly, maker of the Plotly Studio analytics agent and the Dash framework, self-described, Oct 26, 2026. GPT-6.1 Terra through OpenRouter at its default reasoning setting on Plotly's private 43-question data analytics exam.
Favorable
The cheap models keep getting better.
GPT-6.1 Terra. thefourthchime, Hacker News commenter, self-described, Oct 27, 2026. GPT-6.1 Terra on PacBench, a Pac-Man build bake-off, against Grok 4.7.
Favorable
GPT-6.1 Terra is an AWESOME release for anyone using the API in agentic production applications where cost matters.
GPT-6.1 Terra. __jl__, Hacker News commenter, self-described, Oct 26, 2026. GPT-6.1 Terra under internal evals for an agentic production application on the OpenAI API.
Favorable
Terra also raised 27 valid majors against 9, and 4 invalid findings against 0, at 55% of Sol's Codex token count.
Holdout for gpt-6.1-terra against gpt-6.1-sol in revmux, pre-registered. Seigiard, GitHub issue author, self-described, Oct 28, 2026. GPT-6.1 Terra (gpt-6.1-terra) as the Codex reviewer in a 27-round revmux code-review pilot against gpt-6.1-sol.
Favorable
Terra: 6m 22s — $0.29 Sol: 16m 01s — $0.80
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against GPT-6.1 Sol, both at Extra High, on the same car computer use task.
Favorable
Terra: $0.93 — ~21 min Haiku: $2.50 — ~56 min
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against Claude Haiku 5.5 on the same build.
Favorable
it averaged 62.16 (58.47 / 66.34 / 61.68) at $0.22 and 2m 56s a run
X post by @RichDevLab. Richard Thomas (RichDevLab), architect and author of farmerbench.dev per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in Codex at medium effort on three 6-hour FarmerBench seeds.
Unfavorable
2x slower - 16x the cost
X post by @RimasXYZ. Rimas (RimasXYZ), building adrio.ai per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in a blind test on the author's ad prompt agent against the model already in production.
Favorable
6.1-terra is a very interesting model in terms of value for money
X post by @Parad0x80. Parad0x80, CTO per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra in the API and in Codex as a sub-agent.
Unfavorable
I ate through 30% usage yesterday on Terra on my $200/mo plan.
X post by @Maticated. Mat Windle (Maticated), AI developer for work per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra in Codex on the $200 a month plan.

Speed 83%

Favorable
A typical question took 15 seconds , which makes it the fastest model on our leaderboard of 35 to score 40 or more.
GPT-6.1 Terra Benchmark Results for Data Analysis. Plotly, maker of the Plotly Studio analytics agent and the Dash framework, self-described, Oct 26, 2026. GPT-6.1 Terra through OpenRouter at its default reasoning setting on Plotly's private 43-question data analytics exam.
Favorable
in 5.4 min for $0.19 vs Grok's 36.5 min
GPT-6.1 Terra. thefourthchime, Hacker News commenter, self-described, Oct 27, 2026. GPT-6.1 Terra on PacBench, a Pac-Man build bake-off, against Grok 4.7.
Favorable
seems 6.1 terra is indeed fast while I switch from the 6.1 sol when do automated test.
GPT-6.1 Terra. mintflow, Hacker News commenter, self-described, Oct 27, 2026. GPT-6.1 Terra in automated tests after switching from GPT-6.1 Sol.
Favorable
Less than half the time.
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against GPT-6.1 Sol, both at Extra High, on the same car computer use task.
Favorable
Terra was much faster and cheaper.
X post by @AlphaDevj. Alpha (AlphaDevj), developer per the X bio, self-described, Oct 27, 2026. GPT-6.1 Terra against Claude Haiku 5.5 on the same build.
Unfavorable
7/24 runs timed out - 2x slower
X post by @RimasXYZ. Rimas (RimasXYZ), building adrio.ai per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in a blind test on the author's ad prompt agent against the model already in production.

Reliability Only concerns on record

Unfavorable
- 7/24 runs timed out
X post by @RimasXYZ. Rimas (RimasXYZ), building adrio.ai per the X bio, self-described, Oct 26, 2026. GPT-6.1 Terra in a blind test on the author's ad prompt agent against the model already in production.