Kimi K3.1 Pro ranks fourth on Intelligence Index
Moonshot's Kimi K3.1 Pro scored 55 on Artificial Analysis's Intelligence Index, placing fourth overall, and will become the top open-weights model once its weights are released.
Moonshot's new reasoning model, Kimi K3.1 Pro, scores 55 on the Intelligence Index, Artificial Analysis reported. That is 11 points above Kimi K3 (44). Only Claude Opus 5.5 (58), Claude Sonnet 5.5 (56) and Muse Wonder Full (56) score higher.
The weights are not yet public. Moonshot is expected to release them soon after launch. When it does, K3.1 Pro will be the highest-scoring open-weights model, and the index gap between the best open-weights and best proprietary models will drop from 12 points to 3.
Most of the improvement is in agentic tasks. K3.1 Pro has the highest Terminal-Bench 4.0 score of any model at 67%, up from 13% for Kimi K3 and ahead of Claude Sonnet 5.5 at 64%. It ranks third on GDPval-AA with 1,846 Elo, a gain of 312, behind Claude Opus 5.5 (1,866) and Muse Wonder Full (1,857). On AutomationBench-AA, which tests agentic SaaS workflows, it rose from 58% to 73%.
Other evals moved less:
- Humanity's Last Exam: 47% to 52%
- CritPt: 23% to 25%
- SciCode: 59% to 61%
- AA-Omniscience Index: 20 to 30. The hallucination rate fell from 53% to 44%.
Prices are $2 per million input tokens, $10 per million output tokens and $0.20 per million cache-hit tokens, all 33% below Kimi K3. The model uses 14% more output tokens per index task than K3 (55K vs. 48K), so cost per task stays about the same at $1.98.
The model has 2.8T total parameters with 104B active, a 1M-token context window, and accepts text and image input. Artificial Analysis did not say which license the weights will use.