GR
Groq
Serverless Featured
Fastest LLM inference powered by custom LPU chips. OpenAI-compatible API with sub-second latency.
Features
OpenAI Compatible
Streaming
Batching
Fine-tuning
Embeddings
Vision
Audio
Function Calling
JSON Mode
Compliance
SOC 2
HIPAA
GDPR
Pricing Model
per token Free tier: Available
Details
Models 8
API Base
https://api.groq.com/openai/v1 Audio / MusicLLM Inference
Model Catalog (8)
| Model | Type | Input $/1M | Output $/1M | Context | Speed | Status |
|---|---|---|---|---|---|---|
| GPT OSS 120B OpenAI · 117B MoE (5.1B active) | llm | $0.150 | $0.600 | 131k | — | |
| Kimi K2 Moonshot AI · 1T MoE (32B active) | llm | $1.00 | $3.00 | — | — | |
| Llama 3.3 70B Meta · 70B | llm | Enterprise only: Groq lists th | see notes | — | — | |
| Llama 4 Scout Meta · 109B (17B active) | llm | $0.110 | $0.340 | — | — | |
| Qwen 3 32B Alibaba · 32B | llm | $0.290 | $0.590 | — | — | |
| Qwen3.8 27B Alibaba | llm | $0.800 | $4.00 | 131k | — | |
| Whisper Large V3 OpenAI · 1.5B | speech to_text | $0.0000/sec | — | — | — | |
| gpt-oss-20b OpenAI | llm | $0.075 | $0.300 | 131k | — |
Groq API FAQ
Is the Groq API free?
Yes, Groq has a free tier. Paid usage is billed per model.
What is the cheapest Groq model?
Gpt-oss-20b at $0.07 in / $0.30 out per 1M tokens, based on the 8 models we track for Groq.
How many models does Groq offer?
We track 8 Groq models: 7 LLM, 1 speech-to-text.
Is the Groq API OpenAI-compatible?
Yes. Groq exposes an OpenAI-compatible endpoint, so you can use the OpenAI SDKs by changing the base URL and API key.