Kimi K3.1 Pro
Moonshot AI · proprietary · kimi-k3.1-pro
Kimi K3.1 Pro is the first model in Moonshot AI's K3.1 series, a post-trained version of Kimi K3 built with substantially larger post-training compute and large-scale joint reinforcement learning across hundreds of domains. Moonshot describes it as its most capable 3T-class open-weight model, with a 1,048,576-token context window, text/image/video input, tool calling, structured output, and always-on reasoning (low, high, or max; Moonshot API default max). max_completion_tokens defaults to 131,072 and can be raised to 1,048,576. It targets long-horizon coding, 3D creation and CAD, financial research, and other agentic work. Pricing is a one-third cut from K3 ($2 / $10 vs $3 / $15 per MTok). A Highspeed variant (kimi-k3.1-pro-highspeed) offers 2-3x the speed with the same capabilities at $4 / $20 (cached input $0.20). Full weights are scheduled for release by Nov 4, 2026.
Benchmark scores
| Benchmark | Score |
|---|---|
| GDPval-AA 2.1 | 1846.00 |
| AA-Briefcase v1.1 | 1678.00 |
| BabyVision | 0.95 |
| GPQA | 0.94 |
| BenchCAD (with Python tool) | 0.93 |
| BrowseComp | 0.92 |
| Chartography | 0.90 |
| MMMU-Pro (with tools) | 0.90 |
| CharXiv-R | 0.89 |
| MMMU-Pro | 0.84 |
| Kimi Code Bench v2 | 0.77 |
| DeepSWE | 0.76 |
| ExploitBench | 0.75 |
| Humanity's Last Exam | 0.51 |
Pricing
- Moonshot AI$2.00 / $10.00
Input / output per 1M tokens
AA metrics
- Intelligence
- —
- Speed tok/s
- 32.24
- Blended $/M
- —
Arena Elo
- No Arena snapshot linked.
