Supported Models
The models hosted BitRouter routes to natively — a curated SOTA set for agentic and coding work, with input, cached-input, and output pricing — plus how to bring your own.
These are the models hosted BitRouter currently supports. Request a catalog model ID to route among eligible providers and fall back after retryable failures. The hosted catalog focuses on models for agentic and coding workloads.
The self-hosted proxy can also route to your own local or remote models, including fine-tunes and private endpoints: see bring your own model. To use your own provider credentials with the local proxy, see bring your own provider.
Model catalog
| Model | Context | Modalities | Open weights | Input $/M | Cached input $/M | Output $/M |
|---|---|---|---|---|---|---|
anthropic/claude-fable-5 | 1M | text, image | — | $10 | $1 | $50 |
anthropic/claude-fable-5.1 | 1M | file, image, text | — | $10 | $0.25 | $50 |
anthropic/claude-haiku-4.5 | 200K | text, image | — | $1 | $0.1 | $5 |
anthropic/claude-haiku-5.5 | 1M | file, image, text | — | $0.1 | $0.01 | $0.5 |
anthropic/claude-opus-4.8 | 1M | text, image | — | $5 | $0.5 | $25 |
anthropic/claude-opus-5 | 1M | text, image | — | $5 | $0.5 | $25 |
anthropic/claude-opus-5.5 | 1M | file, image, text | — | $4 | $0.2 | $20 |
anthropic/claude-sonnet-4.6 | 1M | text, image | — | $3 | $0.3 | $15 |
anthropic/claude-sonnet-5 | 1M | text, image | — | $2 | $0.2 | $10 |
anthropic/claude-sonnet-5.5 | 1M | file, image, text | — | $2 | $0.2 | $10 |
deepseek/deepseek-v4.1-flash | 1M | image, text | — | $0.65 | $0.13 | $1.45 |
google/gemini-3.7-flash | 1M | audio, file, image, text, video | — | $0.75 | $0.075 | $3.75 |
google/gemini-3.8-flash | 1M | audio, file, image, text, video | — | $0.75 | $0.075 | $3.75 |
google/gemini-4-argon | 1M | audio, file, image, text, video | — | $2 | $0.1 | $10 |
minimax/minimax-m2.7 | 192K | text | ✅ | $0.225 | $0.045 | $0.9 |
minimax/minimax-m3 | 1M | text, image | ✅ | $0.23 | $0.05 | $0.96 |
moonshotai/kimi-k2.7-code | 256K | text, image | ✅ | $0.7125 | $0.1425 | $3 |
moonshotai/kimi-k3 | 1M | text, image | ✅ | $2 | — | $6 |
openai/gpt-5.6-luna | 400K | text, image | — | $0.2 | $0.02 | $1.2 |
openai/gpt-5.6-sol | 1M | text, image | — | $5 | $0.5 | $30 |
openai/gpt-5.6-terra | 1M | text, image | — | $2 | $0.2 | $12 |
openai/gpt-6-astra | 1.1M | file, image, text | — | $10 | $1 | $50 |
openai/gpt-6-luna | 1.1M | file, image, text | — | $0.1 | $0.01 | $0.5 |
openai/gpt-6-sol | 1.1M | file, image, text | — | $2 | $0.2 | $10 |
openai/gpt-6.1-sol | 1.1M | file, image, text | — | $2 | $0.1 | $10 |
qwen/qwen3.7-flash | 1M | image, text, video | — | $0.03 | $0.003 | $0.13 |
qwen/qwen3.7-max | 1M | text | — | $1.875 | $0.375 | $5.625 |
qwen/qwen3.7-plus | 1M | text, image | — | $0.4 | $0.08 | $1.6 |
qwen/qwen3.8-2.4t-a95b | 1M | text | ✅ | $2 | $0.25 | $6 |
qwen/qwen3.8-27b | 1M | text, image, video | ✅ | $0.99 | — | $1.49 |
qwen/qwen3.8-flash | 1M | image, text, video | — | $0.15 | $0.016 | $0.47 |
qwen/qwen3.8-max | 1M | text, image, video | — | $2 | $0.25 | $6 |
qwen/qwen3.8-omni-flash | 1M | audio, image, text, video | — | $0.15 | $0.016 | $0.47 |
x-ai/grok-4.3 | 1M | text, image | — | $1.25 | $0.2 | $2.5 |
x-ai/grok-4.5 | 500K | text, image | — | $2 | $0.3 | $6 |
x-ai/grok-4.6 | 500K | text, image | — | $2 | $0.5 | $6 |
x-ai/grok-4.7 | 500K | file, image, text | — | $2 | $0.5 | $6 |
x-ai/grok-build-0.1 | 256K | text | — | $1 | $0.2 | $2 |
xiaomi/mimo-v2.5 | 256K | text | ✅ | $0.14 | $0.0028 | $0.28 |
xiaomi/mimo-v2.5-pro | 256K | text | ✅ | $0.435 | $0.0036 | $0.87 |
xiaomi/mimo-v2.6-flash | 1M | audio, image, text, video | — | $0.14 | $0.0028 | $0.28 |
xiaomi/mimo-v2.6-pro | 1M | audio, image, text, video | — | $0.435 | $0.0036 | $0.87 |
xiaomi/mimo-v2.6-pro-ultraspeed | 1M | audio, image, text, video | — | $4.35 | $0.036 | $8.7 |
z-ai/glm-5.2 | 1M | text | ✅ | $0.979 | $0.182 | $3.08 |
z-ai/glm-5.3 | 1M | text | — | $1.4 | $0.26 | $4.4 |
z-ai/glm-5.3-flash | 1M | text, image, video | ✅ | $0.075 | $0.015 | $0.25 |
z-ai/glm-5.3-flashx | 1M | image, text, video | — | $0.37 | $0.075 | $1.25 |
Prices are USD per million tokens from the currently available provider with the lowest combined uncached input and text output rates. They are catalog reference rates, not a guarantee of the price paid on every request: the selected provider, cache use, and context tier can change the charge. — means a rate is unavailable in the catalog. Modalities are the inputs a model accepts; every model returns text. The public, open-source registry can also contain providers that are not currently reachable in the hosted service.
How model ids work
On BitRouter a "model" is not a single endpoint. It's an aggregate: one logical model — say openai/gpt-4o or anthropic/claude-sonnet-4.6 — that can be served by many providers at once. You address it by the stable model id in the catalog above, and BitRouter decides which underlying provider endpoint actually answers each request. That indirection is the point: you write your agent against anthropic/claude-sonnet-4.6, and the set of providers behind it can grow, shrink, or re-price without you changing a line of code.
Cloud filters eligible providers by request capabilities and ranks them using price, observed latency, throughput, and health. Hosted routing profiles such as :cost, :latency, and :throughput bias the ranking toward that dimension; they do not guarantee a particular provider. For the self-hosted alpha.33 proxy, model resolution uses configured priority and deterministic fallback order instead of live cost or latency ranking.
Calling these models
The hosted catalog is available through one account at api.bitrouter.ai, without upstream provider keys or per-provider signups. Billing uses the selected route's provider token rates with 0% token markup, and model routing is included. Third-party payment-processing fees may apply when adding funds; those top-up fees are separate from model pricing.
bro cloud login # one-time device-flow sign-in
bro start # the `bitrouter` provider auto-enables once signed inHow is this guide?