Kimi K3
| Maker | Moonshot |
|---|---|
| Family | Kimi |
| Released | 2026-07-16 |
| Context | 1M |
| Max output | 1M |
| Open weights | Kimi K3 License |
|---|---|
| Modalities | text, image, video → text |
| Input / 1M | $3.00 |
| Output / 1M | $15.00 |
| Review call | $0.1149 |
| Reasoning efforts | low, high, max |
|---|---|
| Hosts | Moonshot (Kimi), OpenRouter |
| Model id | kimi-k3 |
CodePanel's benchmark has not run Kimi K3 as a pull-request reviewer yet, so this page has no recall, false-alarm or measured-cost figures for it. The methodology explains how ranked models are scored.
No Kimi model is ranked yet. Of Moonshot's ranked models, Kimi K2.7 Code (Workers AI) is the closest in list price; it caught 38% [28%–47%] of the known bugs at rank 11 of 17, with a measured review cost of $0.0567. At list price, a typical review call (26.2K tokens in and 2.4K out) costs $0.1149 on Kimi K3 and $0.0346 on Kimi K2.7 Code (Workers AI), so this model's call is 3.3× the cost of that one's.
| Host | Model id | Input / 1M | Output / 1M |
|---|---|---|---|
| Moonshot (Kimi) | kimi-k3 | $3.00 | $15.00 |
Source: Moonshot pricing, Oct 16, 2026
| Version | Released | Input / 1M | Output / 1M | Context |
|---|---|---|---|---|
| Kimi K2.6 | 2026-04-20 | $0.95 | $4.00 | 262.1K |
| Model | Maker | Input / 1M | Output / 1M | Review call | Context | Bugs caught |
|---|---|---|---|---|---|---|
| Claude Sonnet 4.6 | Anthropic | $3.00 | $15.00 | $0.1149 | 1M | – |
| GPT-5.4 | OpenAI | $2.50 | $15.00 | $0.1018 | 1.1M | – |
| MiMo-V2.6-Pro UltraSpeed (OpenRouter) | Xiaomi | $4.35 | $8.70 | $0.135 | 1M | – |
| Gemini 3.1 Pro | $2.00 | $12.00 | $0.0814 | 1M | 67% |
Updated Oct 23, 2026