Meta-Models

Don’t pick a provider. Pick a strategy. FreeLLM exposes three meta-models that route across providers based on what you care about.

Model	Strategy	Best For
free	Round-robin across all available providers	Maximum uptime
free-fast	Latency-priority: Groq → Cerebras → Gemini → NIM → Mistral	Real-time chatbots, low latency UIs
free-smart	Capability-priority: Gemini → NIM → Groq → Mistral → Cerebras	Complex reasoning, longer context

Direct provider targeting

You can also target a specific provider model directly:

groq/llama-3.3-70b-versatile
gemini/gemini-2.5-flash
mistral/mistral-small-latest
cerebras/llama3.1-8b
nim/meta/llama-3.3-70b-instruct
nim/nvidia/llama-3.1-nemotron-70b-instruct
nim/deepseek-ai/deepseek-r1

When you target a specific provider model, FreeLLM still applies multi-key rotation, circuit breakers, and rate-limit tracking. You just lose the cross-provider failover.