Available through Groq
openai/gpt-oss-20b
openai/gpt-oss-20b on groq · Groq
What it takes and gives
a prompt → text · temperature
Best fordeclared, not tested
Prompt expansion where latency dominates — it is the fastest route.
Good at
very low latency
Price
About 0.01 credits per run · measured on real runs
Speed and reliability
Fast for a text model
7 runs in 30 days · typically 0.8 s · Groq: 7 runs in 30 days
Also runs on
- anthropic/claude-haiku-4-5 on Opper · ≈ 0.01 credits · sameIn app · own key on API
- claude-haiku-4-5 on Anthropic · ≈ 0.01 credits · same
- deepseek-v4-flash on DeepSeek · ≈ 0.01 credits · sameYour own key
- gemini-3.5-flash-lite on Google Gemini · ≈ 0.01 credits · sameYour own key
- gpt-5.4-nano on OpenAI · ≈ 0.01 credits · same
- mistral-large-latest on Mistral · ≈ 0.01 credits · same
Where it runs
Sushibelt-integratedRuns on Sushibelt's Groq account, billed as credits — or on your own Groq key if your team adds one.
Runs on your own Groq key, or on credits
On your own key, Groq’s terms apply to your team directly.
text-to-text
Workflows
Next stepImages