Model Catalog
base_url.
Every model from Anthropic, OpenAI, Google, x-AI, Moonshot, Perplexity, and DeepSeek —
accessible through OpenAI & Anthropic-compatible endpoints. Set model="auto"
and let Thompson sampling pick the best one for each request.
Add your API keys and BrainstormRouter auto-discovers every available model. No configuration files, no model lists to maintain.
gpt-4o
gpt-4-turbo
o1
o3-mini
gpt-4o-mini
gpt-4.1
gemini-2.5-pro
gemini-2.5-flash
gemini-2.0-flash
gemini-1.5-pro
grok-3
grok-3-mini
grok-2
claude-opus-4
claude-sonnet-4
claude-haiku-3.5
moonshot-v1-8k
moonshot-v1-32k
moonshot-v1-128k
sonar
sonar-pro
sonar-reasoning-pro
deepseek-chat
deepseek-reasoner
model="auto" and forgetThompson sampling learns the optimal model for each task type without manual configuration. Simple tasks route to DeepSeek at $0.000005/req. Complex reasoning routes to Claude Opus at $0.000156/req — a 31x cost difference, chosen automatically.
| Variant | Strategy | Best For |
|---|---|---|
auto |
Thompson Sampling picks based on learned quality posteriors | General workloads |
auto:fast |
Lowest p50 latency model | Real-time UX, streaming |
auto:floor |
Cheapest model above quality threshold | Bulk processing, classification |
auto:best |
Highest quality regardless of cost | Critical reasoning, code review |
Every response from BrainstormRouter includes intelligence headers that tell you which model was selected, why, what it cost, and whether it came from cache.
Rankings derived from production traffic, updated every 5 minutes. Sort by any dimension to find the right model for your workload.
The Pareto frontier shows models where no other option is both cheaper AND higher quality. Points on the line are optimal trade-offs; points below are dominated.
One curl. Auto-routing across 321 models.
curl https://api.brainstormrouter.com/v1/chat/completions \
-H "Authorization: Bearer br_live_..." \
-H "Content-Type: application/json" \
-d '{
"model": "auto",
"messages": [{"role": "user", "content": "Hello"}]
}'