Managed Models
The hosted BitRouter Cloud provider — one account, no upstream keys — the full model catalog with pricing, and automatic discounts on open models.
The BitRouter Cloud provider lets an agent call any model below with a single BitRouter account — no upstream provider keys, no per-provider signups. You pay BitRouter directly at the prices listed here, billed per request; failed requests aren't billed.
bitrouter auth login # one-time device-flow sign-in
bitrouter start # the `bitrouter` provider auto-enables once signed inPrefer your own provider accounts? Use BYOK instead — you pay providers directly at their list price. Running your own model? See local & private models (free).
Supported models & pricing
Prices are USD per million tokens, refreshed at each docs build from the live catalog. Open models are served 25% below official by default — see Discounted open models below.
| Model | Context | Input $/M | Output $/M | Providers |
|---|---|---|---|---|
| anthropic/claude-fable-5 | 1M | $10.00 | $50.00 | 1 |
| anthropic/claude-haiku-4.5 | 200k | $1.00 | $5.00 | 1 |
| anthropic/claude-opus-4.5 | 200k | $5.00 | $25.00 | 1 |
| anthropic/claude-opus-4.6 | 200k | $5.00 | $25.00 | 2 |
| anthropic/claude-opus-4.7 | 200k | $5.00 | $25.00 | 2 |
| anthropic/claude-opus-4.8 | 1M | $5.00 | $25.00 | 1 |
| anthropic/claude-sonnet-4.5 | 200k | $3.00 | $15.00 | 1 |
| anthropic/claude-sonnet-4.6 | 1M | $3.00 | $15.00 | 3 |
| deepseek/deepseek-v3.2 | 131k | $0.21 | $0.32 | 5 |
| deepseek/deepseek-v4-flash | 262k | $0.10 | $0.21 | 5 |
| deepseek/deepseek-v4-pro | 256k | $1.30 | $2.61 | 5 |
| google/gemini-2.5-pro-preview | 1.0M | $1.25 | $10.00 | 1 |
| google/gemini-3-flash-preview | 1.0M | $0.50 | $3.00 | 1 |
| google/gemini-3-pro-preview | 1.0M | $2.00 | $12.00 | 1 |
| google/gemini-3.1-flash-lite-preview | 1M | $0.25 | $1.50 | 1 |
| google/gemini-3.1-pro-preview | 2M | $2.00 | $12.00 | 2 |
| google/gemini-3.5-flash | 1.0M | $1.50 | $9.00 | 1 |
| minimax/minimax-m2.5 | 197k | $0.23 | $0.90 | 6 |
| minimax/minimax-m2.7 | 197k | $0.23 | $0.90 | 3 |
| minimax/minimax-m3 | 1.0M | $0.45 | $1.80 | 2 |
| moonshotai/kimi-k2.5 | 262k | $0.44 | $2.00 | 5 |
| moonshotai/kimi-k2.6 | 256k | $0.71 | $3.00 | 8 |
| moonshotai/kimi-k2.7-code | 262k | $0.71 | $3.00 | 1 |
| openai/gpt-5.1 | 400k | $1.25 | $10.00 | 1 |
| openai/gpt-5.1-codex | 400k | $1.25 | $10.00 | 1 |
| openai/gpt-5.1-codex-mini | 400k | $0.25 | $2.00 | 1 |
| openai/gpt-5.2 | 400k | $1.75 | $14.00 | 1 |
| openai/gpt-5.2-codex | 400k | $1.75 | $14.00 | 1 |
| openai/gpt-5.3-codex | 400k | $1.75 | $14.00 | 1 |
| openai/gpt-5.4 | 128k | $2.50 | $15.00 | 3 |
| openai/gpt-5.4-mini | 128k | $0.75 | $4.50 | 3 |
| openai/gpt-5.4-nano | 400k | $0.20 | $1.25 | 1 |
| openai/gpt-5.5 | 128k | $5.00 | $30.00 | 3 |
| openai/gpt-oss-120b | 131k | $0.25 | $0.69 | 1 |
| qwen/qwen3.6-flash | 1M | $0.19 | $1.13 | 1 |
| qwen/qwen3.6-plus | 131k | $0.38 | $2.25 | 4 |
| qwen/qwen3.7-max | 1M | $1.88 | $5.63 | 3 |
| qwen/qwen3.7-plus | 1M | $0.38 | $2.25 | 2 |
| stepfun/step-3.5-flash | 262k | $0.07 | $0.22 | 2 |
| stepfun/step-3.7-flash | 256k | $0.15 | $0.86 | 2 |
| x-ai/grok-4.1-fast | 131k | $0.20 | $0.50 | 1 |
| x-ai/grok-4.20 | 131k | $1.25 | $2.50 | 2 |
| x-ai/grok-4.3 | 1M | $1.25 | $2.50 | 1 |
| xiaomi/mimo-v2-flash | 66k | $0.09 | $0.29 | 3 |
| xiaomi/mimo-v2-omni | 262k | $0.30 | $1.50 | 2 |
| xiaomi/mimo-v2-pro | 1.0M | $0.75 | $2.25 | 2 |
| xiaomi/mimo-v2.5 | 262k | $0.30 | $1.50 | 3 |
| xiaomi/mimo-v2.5-pro | 262k | $0.75 | $2.25 | 3 |
| z-ai/glm-5 | 203k | $0.75 | $2.40 | 6 |
| z-ai/glm-5.1 | 128k | $1.05 | $3.30 | 7 |
Discounted open models
BitRouter runs its own self-hosted provider for open models, priced 25% below official rates. You get that price automatically — and open-source builders can apply for a deeper custom discount.
25% off by default
Every model except the closed-source families — OpenAI (gpt-*), Anthropic (claude-*), Google (gemini-*), and xAI (grok-*) — is served by BitRouter's self-hosted provider at 25% below the model's official price.
This takes no suffix and no configuration. Because the self-hosted provider is the cheapest source for these models, normal routing already sends your requests there and bills the discounted rate. (The four closed-source families above aren't on the self-hosted provider, so they route to their usual upstreams at standard pricing.)
Pin to the self-hosted provider with :discount
Append :discount to a model id to route the request specifically to BitRouter's self-hosted provider:
curl http://127.0.0.1:4356/v1/chat/completions \
-H "Content-Type: application/json" \
-d '{
"model": "moonshotai/kimi-k2.6:discount",
"messages": [{"role": "user", "content": "Translate to French: Hello."}]
}'The suffix rides on the model string — no body fields, no SDK — and works the same on the OpenAI, Anthropic, and Google surfaces (/v1/messages, /v1beta/models/{model}:generateContent). Use it to guarantee your traffic lands on the discounted self-hosted supply; it's also where any custom discount on your account applies.
Custom discounts up to 50% for open-source projects
Building an open-source agent harness or another open-source project on BitRouter? We offer customized discounts — up to 50% off — for you and your community.
Reach out to set it up:
- Email kelsenliu@bitrouter.ai
- Or book a meeting with the founder:
How is this guide?