Free APILLMAPI keyOfficial
SambaNova Cloud
A truly persistent free tier on RDU hardware, no card, no expiring credits
SambaNova Cloud serves open models (Llama 3.3 70B, Llama 3.1 up to 405B, DeepSeek, Qwen, gpt-oss-120b) on its custom RDU chips at high speed. Unlike credit-based offers, it has a persistent Free Tier that applies automatically whenever no payment method is linked to the account, exposed via an OpenAI-compatible endpoint.
Free tier
Persistent Free Tier with no credit card (applies when no payment method is on file). A separate $5 promotional credit may also be granted that expires in ~30 days, but the rate-limited free tier itself does not expire.Rate limits
Free tier is roughly 10-30 RPM depending on model size (verified June 2026): smaller 8B models near 30 RPM, 405B-class models near 10 RPM. Larger models sit at the lower end.Auth
API keyBase URL
https://api.sambanova.ai/v1What you'd build with it
- A no-expiry free backend for personal projects and Claude Skill fallbacks
- Fast inference on large open models (Llama 405B, Llama 70B) without GPUs
- Low-volume automations that fit inside ~10-30 RPM
- Speed comparisons against GPU-based providers
Key endpoints
- POST
/chat/completionsOpenAI-compatible chat completion (streaming supported) - GET
/modelsList models available to your account
Members
Members get the runnable recipe
Sign in free to copy the example request, see a sample response, and read the gotchas.
Tags
llminferenceopenai-compatiblehigh-throughputopen-weights