// alternatives to

OpenRouter Alternatives

OpenRouter provides one API for 100+ models with some free options, but free model access is limited per day. These providers offer dedicated free tiers with less rotation.

G

Groq

LLM
96
/ 100

Fastest LLM inference in the world. The free tier is real.

14,400 req/day · 30 req/min · no card✓ no card
OpenAI-compatibleStreamingFunction calling
Get API
HF

HuggingFace Inference

LLM
89
/ 100

Run 100,000+ open models. $0.10/month serverless free.

$0.10/month serverless · no card✓ no card
Open sourceHuge catalogEmbeddings
Get API
G

Google Gemini

LLM
95
/ 100

Gemini Flash free forever. 1M context. 1500 req/day.

1,500 req/day · 1M context · no card✓ no card
Vision1M contextMultimodal
Get API
CB

Cerebras

LLM
93
/ 100

2,000+ tokens/sec on Llama 70B. Free tier available.

30 req/min · 1M tokens/day · no card✓ no card
Fastest inferenceOpenAI-compatibleStreaming
Get API
SN

SambaNova

LLM
89
/ 100

RDU-powered inference. Llama 405B for free.

600 req/min · free · no card✓ no card
Fast inferenceLlama 405BOpenAI-compatible
Get API
GH

Glhf.chat

LLM
83
/ 100

Free unlimited Llama 3 405B access. No card ever.

Unlimited · Llama 405B · no card ever✓ no card
Free foreverLlama 405BOpenAI-compatible
Get API
CH

Chutes.ai

LLM
78
/ 100

Decentralized inference on Bittensor. Pay-per-token, some free.

Some free models · $0–$0.30/MTok · no card✓ no card
DecentralizedBittensorUltra-cheap
Get API