Pick a model. Providers fall back in order of free tokens a day.
5 providers with published limits, most tokens first
5 free deployments of GPT OSS 120B, tried in this order: Vercel AI Gateway → Cloudflare Workers AI → Groq → SambaNova → Sakura AI Engine.
# LiteLLM proxy: litellm --config config.yaml
# openai/gpt-oss-120b: 5 free deployments, best first; each falls back to the next.
model_list:
- model_name: openai/gpt-oss-120b
litellm_params:
model: openai/openai/gpt-oss-120b
api_base: https://ai-gateway.vercel.sh/v1
api_key: os.environ/AI_GATEWAY_API_KEY
- model_name: openai/gpt-oss-120b--cloudflare
litellm_params:
model: openai/@cf/openai/gpt-oss-120b
api_base: https://api.cloudflare.com/client/v4/accounts/{account_id}/ai/v1
api_key: os.environ/CLOUDFLARE_API_KEY
rpm: 300
- model_name: openai/gpt-oss-120b--groq
litellm_params:
model: openai/openai/gpt-oss-120b
api_base: https://api.groq.com/openai/v1
api_key: os.environ/GROQ_API_KEY
rpm: 30
tpm: 8000
- model_name: openai/gpt-oss-120b--sakura
litellm_params:
model: openai/gpt-oss-120b
api_base: https://api.ai.sakura.ad.jp/v1
api_key: os.environ/SAKURA_AI_ENGINE_TOKEN
- model_name: openai/gpt-oss-120b--sambanova
litellm_params:
model: openai/gpt-oss-120b
api_base: https://api.sambanova.ai/v1
api_key: os.environ/SAMBANOVA_API_KEY
rpm: 20
router_settings:
num_retries: 0
fallbacks:
- openai/gpt-oss-120b:
- openai/gpt-oss-120b--cloudflare
- openai/gpt-oss-120b--groq
- openai/gpt-oss-120b--sakura
- openai/gpt-oss-120b--sambanova
- openai/gpt-oss-120b--cloudflare:
- openai/gpt-oss-120b--groq
- openai/gpt-oss-120b--sakura
- openai/gpt-oss-120b--sambanova
- openai/gpt-oss-120b--groq:
- openai/gpt-oss-120b--sakura
- openai/gpt-oss-120b--sambanova
- openai/gpt-oss-120b--sakura:
- openai/gpt-oss-120b--sambanova