Reference
Models
The models you can run on Boardwalk's managed lane. Pass a model's id to agent(prompt, { model }) (no key needed). You pay its current source price plus a flat 5% margin, metered as real cost: we route to the cheapest source and pass the lower price through. Bring your own key and tokens cost $0 from us.
| Model | Input / Mtok | Output / Mtok | Context |
|---|---|---|---|
GPT-5.5openai/gpt-5.5 | $5.25 | $31.50 | 1.1M |
Gemini 3 Progoogle/gemini-3.1-pro-preview | $2.10 | $12.60 | 1.0M |
Claude Opus 4.8anthropic/claude-opus-4.8 | $5.25 | $26.25 | 1M |
Claude Sonnet 4.6anthropic/claude-sonnet-4.6 | $3.15 | $15.75 | 1M |
GPT-5.4 miniopenai/gpt-5.4-mini | $0.787 | $4.72 | 400K |
Gemini 3.5 Flashgoogle/gemini-3.5-flash | $1.57 | $9.45 | 1.0M |
Claude Haiku 4.5anthropic/claude-haiku-4.5 | $1.05 | $5.25 | 200K |
Rates as of Jun 17, 2026. Source price + 5% margin, metered as real cost. Rates track provider pricing and can change. Don't want to choose? Pass autoas the model and each call routes to a strong default, billed at that model's rate.