Reference
Models catalog
Every public alias you can pass as the model field. Prices are quoted in USD per 1 million tokens, the same unit OpenAI, Anthropic and Google publish on their pricing pages. Relay billing meters at per-token resolution; amounts are shown in USD while your wallet is debited in CN¥ at the equivalent amount.
17 models live. You can also browse this list in the developer console.
claude
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
claude-3-5-sonnet | 💬 Text | $3.00 | $15.00 |
deepseek-v
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
deepseek-v4-flash | 💬 Text | $0.140 | $0.281 |
deepseek-v4-pro | 💬 Text | $0.435 | $0.869 |
doubao-seedance
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
doubao-seedance-2-0-pro | 🎬 Video | — | $6.40 |
gpt
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
gpt-4.1-mini | 💬 Text | $0.150 | $0.600 |
gpt-5.2 | 💬 Text | $1.75 | $14.00 |
gpt-5.3-codex | 💬 Text | $1.75 | $14.00 |
gpt-5.3-codex-spark | 💬 Text | — | — |
gpt-5.4 | 💬 Text | $2.50 | $15.00 |
gpt-5.4-mini | 💬 Text | $0.700 | $4.20 |
gpt-5.4-openai-compact | 💬 Text | $2.50 | $15.00 |
gpt-5.4-xhigh-fast | 💬 Text | $2.50 | $15.00 |
gpt-5.5 | 💬 Text | $5.00 | $30.00 |
gpt-5.5-openai-compact | 💬 Text | $5.00 | $30.00 |
gpt-5.5-xhigh-fast | 💬 Text | $5.00 | $30.00 |
gpt-image
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
gpt-image-2 | 🖼 Image | — | — |
text-embedding
| Alias | Modality | Input / 1M ($) | Output / 1M ($) |
|---|
text-embedding-3-small | 🔢 Embeddings | $0.019 | — |
Routing & failover
You always pass the platform alias (e.g. gpt-5.2), never the upstream-specific id (e.g. openai/gpt-5.2). Behind the scenes:
- Each alias maps to one or more upstream routes ordered by priority.
- Each route consumes credentials from a per-provider pool. If a credential gets repeatedly rejected, it's cooled down for 60s and the next request picks a sibling.
- A single retryable upstream error (5xx, 401/403 token issues, 429 rate limit) triggers the relay to retry the next candidate automatically — your client only sees the final result.
- If every candidate is exhausted, you get an explicit
503 with a clear message instead of a stuck request.
See Common errors for what each status code means and how to react.