Groq API
Groq
Free tier: 30 RPM per organization; daily request caps vary by model — roughly 14,400/day for Llama 3.1 8B Instant, and around 1,000/day for larger models like GPT-OSS 120B/20B and Llama 3.3 70B
TPM is roughly 6,000-8,000 depending on the model. Free long-term, no credit card required; check console.groq.com/settings/limits after logging in for exact figures.
Llama 3.3 70B VersatileLlama 3.1 8B InstantGPT-OSS 120BGPT-OSS 20BWhisper Large v3
Known for fast inference on LPU hardware, suited to latency-sensitive use cases; daily quotas for larger models are noticeably lower than for smaller ones, so heavy use hits the cap quickly.