# Pricing

The models we serve and what they cost.


We serve a short list of models, most of them open-weight, plus a few from Anthropic and OpenAI. If you need one that isn't listed, [tell us](/support).

## Plans

| Plan | Price | Includes | Limits |
| --- | --- | --- | --- |
| Team on-demand | Prepaid credits; no subscription fee | OpenAI- and Anthropic-compatible inference, exact per-request receipts, shared team credits, per-key budgets, and discounted `1h` and `24h` batch windows on the models we serve ourselves (launching soon) | Your team's current throughput cap is shown in [Usage](/usage). |
| Anonymous agent trial | `$0.01` total credit | A temporary key for testing; [mint it using the quickstart](/docs/quickstart) | Expires after 24 hours; 6 requests and 30,000 tokens per minute. It cannot be topped up. |

## Rate card

Per 1M tokens. Cached input is charged at the cached rate when a request reuses a prefix the gateway already holds.

`asap` is today's on-demand rate, and every model has one. `1h` and `24h` are batch windows: upload a file of requests, pick a deadline, and pay 25% or 50% less per token. Only the models we run on our own fleet have them, so a model listed with just `asap` does not. Batch windows launch in the coming weeks.

USD per 1M tokens. `asap` is on-demand. `1h` and `24h` are batch completion windows, launching in the coming weeks on the models we serve ourselves; a model listed only as `asap` does not offer them.

| Model | Lab | Window | Input | Cached input | Output |
| --- | --- | --- | --- | --- | --- |
| `claude-opus-5.5` | Anthropic | `asap` | $4.00 | $0.20 | $20.00 |
| `claude-opus-5` | Anthropic | `asap` | $5.00 | $0.50 | $25.00 |
| `claude-fable-5.1` | Anthropic | `asap` | $10.00 | $0.25 | $50.00 |
| `dsv4` | DeepSeek | `asap` | $0.04 | $0.02 | $0.21 |
| `dsv4` | DeepSeek | `1h` | $0.03 | $0.02 | $0.16 |
| `dsv4` | DeepSeek | `24h` | $0.02 | $0.01 | $0.11 |
| `deepseek-v4.1-flash` | DeepSeek | `asap` | $0.20 | $0.01 | $1.00 |
| `deepseek-v4-pro-0813` | DeepSeek | `asap` | $1.04 | $0.04 | $2.08 |
| `minimax-m3-free` | MiniMax | `asap` | $0 | $0 | $0 |
| `minimax-m3` | MiniMax | `asap` | $0.28 | $0.06 | $1.10 |
| `kimi-k3` | Moonshot AI | `asap` | $2.30 | $0.23 | $11.40 |
| `nemotron3-super` | NVIDIA | `asap` | $0.12 | $0.06 | $0.52 |
| `nemotron3-super` | NVIDIA | `1h` | $0.09 | $0.05 | $0.39 |
| `nemotron3-super` | NVIDIA | `24h` | $0.06 | $0.03 | $0.26 |
| `gpt-6-luna` | OpenAI | `asap` | $0.10 | $0.01 | $0.50 |
| `gpt-5.6-luna` | OpenAI | `asap` | $0.20 | $0.02 | $1.20 |
| `gpt-5.6-sol` | OpenAI | `asap` | $2.00 | $0.20 | $10.00 |
| `gpt-6-sol` | OpenAI | `asap` | $2.00 | $0.20 | $10.00 |
| `qwen3.6` | Qwen | `asap` | $0.04 | $0.02 | $0.35 |
| `qwen3.6` | Qwen | `1h` | $0.03 | $0.02 | $0.26 |
| `qwen3.6` | Qwen | `24h` | $0.02 | $0.01 | $0.18 |
| `qwen3.8-flash` | Qwen | `asap` | $0.15 | $0.02 | $0.47 |
| `mimo-v2.6-flash` | Xiaomi | `asap` | $0.14 | $0 | $0.28 |
| `mimo-v2.6-pro` | Xiaomi | `asap` | $0.44 | $0 | $0.87 |
| `glm-5.3-flash` | Z.ai | `asap` | $0.12 | $0.02 | $0.40 |
| `glm-5.2` | Z.ai | `asap` | $0.34 | $0.17 | $2.30 |
| `glm-5.3` | Z.ai | `asap` | $1.12 | $0.14 | $3.52 |