Comparisons

Both list gpt-oss-120b at US$0.15 input and US$0.60 output per 1M tokens, so GILD prices 100M input and 20M output tokens a month at US$27 on either. Fireworks lists a US$0.015 cached-input rate that Groq's models page does not, and keeps no data for open models by default, and Groq's batch runs at 50% lower cost within 24 hours to 7 days.

Both list gpt-oss-120b at US$0.15 input and US$0.60 output per 1M tokens. Fireworks has Zero Data Retention by default and bills dedicated GPUs per GPU second, while Together AI supports zero data retention at the organization level, off by default, and bills dedicated inference per minute while a deployment runs.