Kimi K3.1 Pro
The strongest open-weight model, within striking distance of the frontier
Overview
Kimi K3.1 Pro is the highest-scoring open-weight model on the Artificial Analysis index in November 2026, closing to within a few points of the closed frontier. Weights are available under a conditional licence, and the model is notable for agentic coding and long-context work — though self-hosting requires substantial hardware.
Best for
- Open-weight deployment
- Agentic coding
- Long documents
Strengths & limitations
✓ Strengths
- Top open-weight intelligence
- Strong Terminal-Bench
- 1M context
△ Limitations
- Conditional licence
- Very large self-host requirement
Reported benchmarks
| Benchmark | Score | Note |
|---|---|---|
| Artificial Analysis Intelligence Index | ~55 | — |
| Terminal-Bench | within ~1pt of Opus 5.5 | — |
Benchmark figures are reported values from public sources and may use different evaluation harnesses. See sources below.
Frequently asked questions
How much does Kimi K3.1 Pro cost?
The API price is $2 per 1M input tokens and $10 per 1M output tokens (as of 2026-11-13). Prices change — check the official provider page for current rates.
Is Kimi K3.1 Pro open source?
Kimi K3.1 Pro is open weights. Weights are available, subject to the licence terms — review the model card before commercial use.
What is the context window of Kimi K3.1 Pro?
Kimi K3.1 Pro supports a context window of 1M tokens with a maximum output of 131072 tokens per response.
What is Kimi K3.1 Pro best at?
Based on our scoring, Kimi K3.1 Pro is strongest at top open-weight intelligence, strong terminal-bench, 1m context. It is a common choice for open-weight deployment, agentic coding, long documents.