Models. The gateway's catalog, with the rate billed right now.
Send the model id exactly as listed, or send auto and let the router pick the cheapest model expected to handle the request.
The catalog. USD per 1M tokens, billed from a credit balance.
| Model | Input | Output | Context | Tools | Speed |
|---|---|---|---|---|---|
Auto (smart routing)autoSmart routing: picks the cheapest capable model and tries the next one up on failure | Billed at the rate of the model that served it | ||||
Qwen 3.8 Maxqwen/qwen3.8-maxFlagship reasoning & coding model, 1M context | $2.40 | $7.20 | 1M | Yes | Medium |
Qwen 3.8 Flashqwen/qwen3.8-flashFast reasoning model, 1M context | $0.18 | $0.564 | 1M | Yes | Fast |
MiMo V2.6-Proxiaomi/mimo-v2.6-proFlagship omnimodal agent model | $0.522 | $1.04 | 262k | Yes | Medium |
DeepSeek V4.1 Flashdeepseek/deepseek-v4.1-flashFast, very low-cost reasoning model, 1M context | $0.0612 | $0.68 | 1M | Yes | Fast |
Claude Opus 5.5anthropic/claude-opus-5.5Anthropic's most capable model, 1M context | $4.80 | $24.00 | 1M | Yes | Slow |
Claude Sonnet 5.5anthropic/claude-sonnet-5.5Frontier coding & agents, 1M context | $2.40 | $12.00 | 1M | Yes | Medium |
Claude Haiku 5.5anthropic/claude-haiku-5.5Fast, low-cost Claude, 1M context | $0.12 | $0.60 | 1M | Yes | Fast |
GPT-6.1 Solopenai/gpt-6.1-solOpenAI's flagship reasoning model, 1M context | $2.40 | $12.00 | 1M | Yes | Medium |
GPT-6.1 Terraopenai/gpt-6.1-terraBalanced reasoning model, 1M context | $1.20 | $6.00 | 1M | Yes | Medium |
GPT-6 Lunaopenai/gpt-6-lunaFast, low-cost GPT, 1M context | $0.12 | $0.60 | 1M | Yes | Fast |
MiMo V2.6-Flashxiaomi/mimo-v2.6-flashFast omnimodal model | $0.168 | $0.336 | 262k | Yes | Fast |
Gemini 3.5 Flash Litegoogle/gemini-3.5-flash-liteLow-latency multimodal model, 1M context | $0.36 | $3.00 | 1M | Yes | Fast |
MiniMax M3minimax/minimax-m3Agentic reasoning model | $0.36 | $1.44 | 205k | Yes | Medium |
Qwen 3.7 Maxqwen/qwen3.7-maxFlagship coding model | $1.77 | $5.31 | 262k | Yes | Slow |
Kimi K3.1 Promoonshotai/kimi-k3.1-proFrontier multimodal reasoning, 1M context | $2.40 | $12.00 | 1M | Yes | Slow |
GLM 5.3z-ai/glm-5.3Agentic coding & reasoning, 1M context | $1.68 | $5.28 | 1M | Yes | Slow |
Nemotron 3.5 Super VL (free)nvidia/nemotron-3.5-super-vl:freeVision-language reasoning MoE, 1M context, free (rate limited)No credits needed | $0 | $0 | 1M | Yes | Medium |
Nemotron 3.5 Super VLnvidia/nemotron-3.5-super-vlVision-language reasoning MoE, 1M context | $0.18 | $0.60 | 1M | Yes | Medium |
Ling 3.1 Flashinclusionai/ling-3.1-flashToken-efficient agentic MoE, 1M context | $0.09 | $0.27 | 1M | Yes | Fast |
Tencent HY4tencent/hy4Multimodal model, 1M context | $0.50 | $1.50 | 1M | Yes | Medium |
Macaron V1 Tallmindai/macaron-v1-tallRouted MoL personal-intelligence model, 262k context | $0.54 | $3.12 | 262k | Yes | Fast |
Macaron V1 Ventimindai/macaron-v1-ventiFlagship routed MoL on GLM-5.2, 1M context | $1.80 | $5.40 | 1M | Yes | Slow |
Live from the gateway's model list, GET /v1/models, re-read every five minutes. The rate billed from a credit balance right now, promotions included.
Retired. Older versions, served as their successor.
| Model | Retired | Requests are served as |
|---|---|---|
GLM 5.2z-ai/glm-5.2 | 2026-11-09 | z-ai/glm-5.3 |
Kimi K3moonshotai/kimi-k3 | 2026-11-09 | moonshotai/kimi-k3.1-pro |
Tencent HY3tencent/hy3 | 2026-11-09 | tencent/hy4 |
Ling 3.0 Flashinclusionai/ling-3.0-flash | 2026-11-09 | inclusionai/ling-3.1-flash |
Ling 3.0 Tiny (free)inclusionai/ling-3.0-tiny:free | 2026-08-13 | inclusionai/ling-3.1-flash |
Gemini 3.1 Flash Litegoogle/gemini-3.1-flash-lite | 2026-11-09 | google/gemini-3.5-flash-lite |
Nemotron 3 Ultra (free)nvidia/nemotron-3-ultra-550b-a55b:free | 2026-11-09 | nvidia/nemotron-3.5-super-vl:free |
Nemotron 3 Ultranvidia/nemotron-3-ultra-550b-a55b | 2026-11-09 | nvidia/nemotron-3.5-super-vl |
MiMo V2.5-Proxiaomi/mimo-v2.5-pro | 2026-10-21 | xiaomi/mimo-v2.6-pro |
MiMo V2.5xiaomi/mimo-v2.5 | 2026-10-21 | xiaomi/mimo-v2.6-flash |
Old ids keep working: the gateway serves them as the newer model and bills at its rate, so existing configs need no change.