Skip to main content

Supported Providers and Models

Routor connects to 13 providers out of the box. All routing happens automatically. You do not choose a provider - Routor picks based on the request.
📷 [Screenshot: Provider health dashboard showing all 13 providers with green status indicators]

Providers

ProviderModels availableNotes
OpenAIGPT-5.4, GPT-5.5Strong general and coding
AnthropicClaude Opus 4.6 / 4.7 / 4.8, Sonnet 4.5 / 4.6, Haiku 4.5Best creative writing and long context
GoogleGemini 3.1 Flash-Lite, Gemini 3.1 Pro, Gemini 3.5 Flash, Gemma 4Fast and cheap at scale
DeepSeekDeepSeek V4 Flash, DeepSeek V4 ProBest reasoning-per-dollar
xAIGrok 4, Grok 4.1 Fast ReasoningStrong general model
NVIDIANemotron 3 UltraFast open-weight inference
MoonshotKimi K2.6, Kimi K2.7 CodeLong context, strong at code
MiniMaxMiniMax M3Strong value across most tasks
Z.AIGLM-5.2, GLM-4.7 FlashOpen-weight, competitive with frontier
Alibaba (Qwen)Qwen3.5-397B, Qwen3.6-27B, Qwen3.7-MaxStrong multilingual
MistralMistral Large 3, Medium 3.5, Small 4, Devstral 2Strong European-hosted option
MicrosoftPhi-4, Mai-Code-1-FlashEnterprise / compliance-friendly routing
XiaomiMiMo-V2.5-ProEfficient bilingual reasoning

Model Pricing

Input and output costs at provider list price. Routor adds a 10% routing margin - see Plans & Pricing.
ModelTierInput /1M tokensOutput /1M tokensContext
microsoft/phi-4NANO$0.10$0.10128K
zai/glm-4.7-flashNANO$0.20$0.20128K
google/gemini-3.1-flash-lite-previewNANO$0.25$1.501M
qwen/qwen3.6-27bSIMPLE$0.30$1.50128K
anthropic/claude-haiku-4.5SIMPLE$1.00$5.00200K
zai/glm-5.2SIMPLE$1.40$4.40128K
deepseek/deepseek-v4-flashLIGHT$0.14$0.28128K
moonshot/kimi-k2.6LIGHT$0.68$3.42256K
google/gemini-3.5-flashLIGHT$1.50$9.001M
mistral/mistral-large-3STANDARD$2.00$6.00128K
openai/gpt-5.4STANDARD$2.50$15.00200K
anthropic/claude-sonnet-4.6STANDARD$3.00$15.00200K
deepseek/deepseek-v4-proCOMPLEX$0.44$0.87128K
xai/grok-4COMPLEX$3.00$15.00131K
anthropic/claude-opus-4.8COMPLEX$5.00$25.00200K
openai/gpt-5.5COMPLEX$5.00$30.00200K
Prices as of June 2026. Model pricing changes frequently - check the Models page in your dashboard for current rates.

Model Capabilities

Not every model supports every capability. Routor filters the pool automatically based on what your request needs.
Provider / ModelVisionFilesToolsAudioVideo
OpenAI GPT-5.5YesYesYesYesNo
OpenAI GPT-5.4YesNoYesNoNo
Claude Opus 4.8YesYesYesNoNo
Claude Sonnet 4.6YesYesYesNoNo
Claude Haiku 4.5YesNoYesNoNo
Gemini 3.1 ProYesYesYesYesYes
Gemini 3.5 FlashYesYesYesYesNo
DeepSeek V4 ProNoNoNoNoNo
GLM-5.2YesNoYesNoNo
Qwen3.7-MaxNoNoYesNoNo

How Providers Are Validated

When Routor starts, it pings every configured provider using their /models endpoint. No tokens consumed, no cost. Any provider that fails because of an invalid key, network error, or rate limit gets removed from routing until the next restart. You will never get routed to a broken provider. Check the current status of all providers in the Provider Health dashboard.

Adding a New Provider

Routor’s admin connects provider API keys on the server side. End users do not manage provider keys - Routor is a managed service, not a self-hosted deployment. You use a Routor API key and Routor handles provider authentication and every underlying provider relationship for you. Only providers with a working key on Routor’s side are included in routing, and that pool grows without any action needed from you.