Inference cloud
How GroqCloud compares
LPU inference. Token prices and tokens/s from the GroqDocs models page. Llama rows there say Contact Sales.
Pricing page · Site · Open in compare
home.cloud.groq.io/pricing listed dollar prices for Llama 3.1 8B and Llama 3.3 70B. console.groq.com/docs/models lists those production models as Contact Sales. The docs page is the source used here, so those Llama prices are empty.
Regions and compliance
Regions: not published / unverified
Compliance: not published / unverified
Token offerings
GPT OSS 120B
serverless_api · openai/gpt-oss-120b · open-weight true
- Input / 1M
- 0.15console.groq.comchecked 2026-10-20“GPT OSS 120B openai/gpt-oss-120b | $0.15 input”
- Cached input / 1M
- not published / unverifiedconsole.groq.comchecked 2026-10-20“Cached input price not published.”
- Output / 1M
- 0.6console.groq.comchecked 2026-10-20“GPT OSS 120B openai/gpt-oss-120b | $0.60 output”
Context 131,072 · tokens/s 500 · TTFT not published
Rate limits are the developer plan column. Tokens/s is the published speed, not a GPU benchmark.
GPT OSS 20B
serverless_api · openai/gpt-oss-20b · open-weight not published / unverified
- Input / 1M
- 0.075console.groq.comchecked 2026-10-20“GPT OSS 20B openai/gpt-oss-20b | $0.075 input”
- Cached input / 1M
- not published / unverifiedconsole.groq.comchecked 2026-10-20“Cached input price not published.”
- Output / 1M
- 0.3console.groq.comchecked 2026-10-20“GPT OSS 20B openai/gpt-oss-20b | $0.30 output”
Context 131,072 · tokens/s 1000 · TTFT not published
Llama 3.3 70B
serverless_api · llama-3.3-70b-versatile · open-weight not published / unverified
- Input / 1M
- not published / unverifiedconsole.groq.comchecked 2026-10-20“Docs list this production model as Contact Sales. A different Groq marketing page showed a dollar price and was not used.”
- Cached input / 1M
- not published / unverifiedconsole.groq.comchecked 2026-10-20“Cached input price not published.”
- Output / 1M
- not published / unverifiedconsole.groq.comchecked 2026-10-20“Docs list this production model as Contact Sales.”
Context 131,072 · tokens/s 280 · TTFT not published
Compute SKUs
No GPU or node price is seeded.