Thinking Kimi model for slower research passes, planning, and hard technical questionsKimi K2 Thinking by moonshotai is offered by 17 providers on this page. Public prices are shown for 16 of them. No official API price is listed for it; the prices below come from third-party channels. The lowest paid channel is Helicone at $0.48 / $2.00 per 1M. That channel is a third-party gateway, so confirm its availability and rate limits before depending on it. Prices vary about 1.6× across the providers that host it.
Median output speed is 137 tokens per second, with 1.45s to the first token.
The context window is 262,144 tokens at the reference host, but hosts report different limits, from 32,768 to 262,144, so the usable window depends on the provider you pick. It supports reasoning, tool use, and structured output. The weights are open, so it can also be self-hosted or served through a gateway of your choice. Its training knowledge cuts off at 2024-08.
Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.
| Provider | Tier | Input | Output | Cache read | Cache write | Context | Output limit | Status |
|---|---|---|---|---|---|---|---|---|
| Helicone | Gateway | $0.48 | $2.00 | — | — | 256,000 ⚠ | 256,000 | |
| OpenCode | Gateway | $0.40 | $2.50 | $0.40 | — | 262,144 | 262,144 | deprecated |
| Alibaba (China) | First-party | $0.57 | $2.29 | — | — | 262,144 | 16,384 | deprecated |
| 302.AI | Gateway | $0.58 | $2.30 | — | — | 262,144 | 262,144 | |
| Qiniu | Gateway | $0.58 | $2.32 | — | — | 256,000 ⚠ | 100,000 | deprecated |
| NanoGPT | Gateway | $0.60 | $2.50 | $0.15 | — | 262,144 | 98,304 | |
| Amazon Bedrock moonshot.kimi-k2-thinking | Cloud | $0.60 | $2.50 | — | — | 262,144 | 16,000 | deprecated |
| Novita AI | Gateway | $0.60 | $2.50 | $0.15 | — | 262,144 | 98,304 | |
| OpenRouter | Gateway | $0.60 | $2.50 | $0.15 | — | 262,144 | 98,304 | |
| Eden AI amazon/moonshot.kimi-k2-thinking | Gateway | $0.60 | $2.50 | — | — | 256,000 ⚠ | 256,000 |
Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.
Interleaved thinking (reasoning between tool calls) is declared by 8 of 17 providers.
14 / 17 providers expose no reasoning control (reasoning_options: []).
| Name | Conditions | Score | Metric | Source |
|---|---|---|---|---|
| SWE-Bench Verified | — | 71.3 | resolved | Source ↗ |
The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.