Skip to content
Kimi K3.1 Pro Launches with Enhanced Capabilities

Kimi K3.1 Pro Launches with Enhanced Capabilities

First seen 23 Oct 2026, 21:14 UTC •

Article Content

Browse articles
ThreatCluster AI
ThreatCluster •October 24, 2026 at 06:13 UTC
  • •Kimi K3.1 Pro features a 1-million-token context window.
  • •The model shows significant performance improvements in various domains.
  • •K3.1 Pro scored 65.7% on Terminal-Bench 4.0, competing closely with top models.

Kimi has launched the K3.1 Pro, a 3T-class open-weight model featuring a 1-million-token context window. The model is designed for improved performance in coding, 3D creation, and financial research. Alongside, the K3.1 Pro Highspeed variant promises 2-3 times the speed of the standard model. Both versions are available on Kimi's platforms, with full model weights expected by November 4. The K3.1 Pro is a post-trained version of Kimi K3, optimized for large-scale reinforcement learning across various domains. In benchmark tests, it scored 65.7% on Terminal-Bench 4.0, outperforming GPT-6 Astra but slightly trailing behind Claude Opus 5.5. The improvements are attributed to optimizations in the RL framework and training efficiency.

Start a free Starter trial for enhanced analysis

Ask AI about this cluster

Updated just now How this analysis works

Timeline

2026-10-23
Kimi K3.1 Pro launched
Kimi introduced the K3.1 Pro and K3.1 Pro Highspeed models, enhancing capabilities across multiple domains.
Kimi.com
2026-10-23
K3.1 Pro performance benchmarked
K3.1 Pro scored 65.7% on Terminal-Bench 4.0, outperforming GPT-6 Astra and closely trailing Claude Opus 5.5.
Kimi.ai

More articles in this cluster (2)

Common questions

What improvements does K3.1 Pro offer?
K3.1 Pro features a 1-million-token context window and enhanced performance in coding, 3D creation, and financial research.
When will the full model weights be available?
The full model weights for K3.1 Pro are expected to be released by November 4.
How does K3.1 Pro compare to other models?
In Terminal-Bench 4.0, K3.1 Pro scored 65.7%, outperforming GPT-6 Astra and closely trailing Claude Opus 5.5.