Model catalog

Every model you can route to. Use an auto/* alias or pin an exact id. The catalog updates itself from provider APIs.

Aliases

Churn-proof shortcuts — they resolve to the best model under your policy.

auto balancedauto/agents balancedauto/chat balancedauto/cheap cheapestauto/cheap-us cheapestauto/code balancedauto/frontier premiumauto/smart balancedauto/us-open balancedauto/vision balanced
ModelClassContextProvidersProvider quality (beta)Capabilities
anthropic/claude-opus-4.8
Claude Opus 4.8
frontier250K2—
tools vision
anthropic/claude-sonnet-4.6
Claude Sonnet 4.6
frontier250K2—
tools vision
google/gemini-2.5-pro
Gemini 2.5 Pro
frontier1000K2—
tools vision
google/gemini-3-flash-preview
Gemini 3 Flash (preview)
frontier1000K2—
tools vision
mistralai/mistral-large
Mistral Large
frontier128K2—
tools
openai/gpt-5.2
GPT-5.2
frontier256K2—
tools vision
x-ai/grok-4
Grok 4
frontier256K2—
tools vision
x-ai/grok-code-fast
Grok Code Fast
frontier256K1—
tools
deepseek/deepseek-v4-flash
DeepSeek V4 Flash
open_weight164K4—
open tools
deepseek/deepseek-v4-pro
DeepSeek V4 Pro
open_weight1000K3—
open tools
meta-llama/llama-3.3-70b
Llama 3.3 70B
open_weight128K2—
open tools
minimax/minimax-m3
MiniMax M3
open_weight256K2—
open tools
moonshotai/kimi-k2
Kimi K2
open_weight256K1—
open tools
moonshotai/kimi-k2.6
Kimi K2.6
open_weight256K3—
open tools
nvidia/nemotron-3-super
Nemotron 3 Super
open_weight128K2—
open tools
nvidia/nemotron-3-ultra
Nemotron 3 Ultra
open_weight1000K2—
open tools
nvidia/nemotron-4
Nemotron 4
open_weight128K1—
open tools
openai/gpt-oss-120b
GPT-OSS 120B
open_weight128K2—
open tools
qwen/qwen3.7-plus
Qwen3.7 Plus
open_weight1000K2—
open tools
qwen/qwen3-max
Qwen3 Max
open_weight256K2—
open tools
1–20 of 30