Groq pricing overview
TokenWatch tracks 6 text, 0 image, and 0 video model records for this provider. TokenWatch does not have a provider-wide zero-retention verdict for this page.
Cheapest input price: Llama 3.1 8B Instruct at $0.050 per million input tokens and $0.080 per million output tokens. 5 of 6 text models publish a cache-read price; 6 carry a zero-data-retention tag.
Groq text-model pricing
Text offerings ranked by the Agentic workload mix (2.5% input, 97% cached input, 0.5% output). Change the calculator mix for a workload-specific estimate.
| Org | Provider | Model | Input $/M | Output $/M | Cache $/M | Blended $/M |
|---|---|---|---|---|---|---|
| meta | groq | Meta: Llama 3.1 8B Instruct | $0.050 | $0.080 | $0.025 | $0.026 |
| openai | groq | OpenAI: gpt-oss-safeguard-20b | $0.075 | $0.300 | $0.037 | $0.040 |
| openai | groq | OpenAI: gpt-oss-20b | $0.075 | $0.300 | $0.037 | $0.040 |
| openai | groq | OpenAI: gpt-oss-120b | $0.150 | $0.600 | $0.075 | $0.080 |
| meta | groq | Meta: Llama 3.3 70B Instruct | $0.590 | $0.790 | $0.295 | $0.305 |
| minimax | groq | MiniMax: MiniMax M2.7 | $0.600 | $1.80 | — | $0.606 |
Catalog snapshot 2026-10-29 from direct-provider APIs, OpenRouter endpoint data, and maintained fallbacks. See the sources and calculation rules; verify rates on the provider's official pricing page before committing spend.