StoryFirst seen

Moonshot removed cyber training from Kimi K3.1 Pro before open-sourcing it

The developer says it will not release open weights that extend the frontier of offensive cyber capability.

Archived Updated

Everything it touches

11 names 6 voices 8 source posts

The story’s neighborhood

NAMESVOICESPOSTS

Explore the graph, or choose a connection below.

Follow the evidence

The source posts

This brief comes from Tech Twitter’s edited headlines. The posts below are the sources grouped with the story; statements in them belong to their authors.

MTSLead source
SITUATION EXPLAINED: China's Kimi K3.1 Pro got so good at hacking that Moonshot stripped its cyber skills before open-sourcing it. • Kimi K3.1 Pro has 2.8T parameters (104B active) and a 1M-token context window, with full weights due by Nov 4 • It nearly matches Opus 5.5 on Terminal-Bench 4.0 (65.7% vs 66.4%) and beats it on SWE Marathon, HLE with tools, and OfficeQA Pro v2 • API pricing is $2/$10 per million tokens, a third cheaper than Kimi K3 • Removing cyber data from training dropped its score on real exploit tasks from 90.5% to 81%, now below DeepSeek V4.1 Flash at 85.7% • Moonshot pledged not to release open weights that "extend the absolute frontier of cyber capability" • It still plans to raise the cyber capability of its open weights over time, saying defenders need dual-use…
Read original on X
Kimi.ai
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%. 🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks. Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi Work, Kimi Code, and the Kimi API. Open weights by Nov 4, 2026. 🔗 API: 🔗 Tech blog:
Read original on X
Read 4 more source posts
ℏεsam
WAIT… Kimi K3.1 Pro became so good at exploiting vulnerabilities that Moonshot had to NERF its hacking capabilities before releasing the weights. > removed cyber data from training and RL > exploit success rate went from 90.5% to 81% > it’s still more than DOUBLE its predecessor’s 33.3% but they actually plan to INCREASE the cyber capabilities in future models. from the blog: “First, we aim not to release open weights that would extend the absolute frontier of cyber capability. K3.1 Pro's offensive cyber capability is, by design, below what our pipeline can produce, and remains behind that of frontier proprietary models. Second, we intend to progressively raise the cyber capability available in open weights. We believe making K3.1 Pro available in a non-discriminatory way benefits people…
Read original on X
OpenRouter
Kimi K3.1 Pro from @Kimi_Moonshot is live on OpenRouter. A 1M context, multimodal open-weight model, post-trained from Kimi K3 for long-horizon coding and agentic work. $2/M input tokens, $10/M output tokens. Try it:
Read original on X

How the headline evolved

  1. Moonshot removed cyber training from Kimi K3.1 Pro before open-sourcing it

  2. Moonshot removes offensive cyber skills from its open-weight Kimi model

  3. Moonshot removes offensive cyber skills from its open-weight Kimi model

  4. Moonshot strips offensive cyber capability from open-weight Kimi model

  5. Moonshot strips offensive cyber capability from its open-weight Kimi model

  6. Moonshot removes cyber training from Kimi model before open release

  7. Moonshot releases open-weight Kimi model after removing cyber training

  8. Moonshot releases open-weight Kimi model with 1M-token context

  9. Moonshot releases open-weight Kimi model with 1M-token context

  10. Moonshot releases open-weight Kimi model with 1M-token context

  11. Moonshot launches open-weight Kimi model with 1M-token context

  12. Moonshot launches open-weight Kimi model with 1M-token context

  13. Moonshot launches open-weight Kimi model with 1M-token context

  14. Moonshot launches open-weight Kimi model with 1M-token context

  15. Moonshot AI launches open-weight Kimi K3.1 Pro for coding and finance

  16. Moonshot AI launches open-weight Kimi K3.1 Pro for coding and finance work

  17. Moonshot AI launches open-weight Kimi model for coding and finance work

  18. Moonshot AI launches Kimi open-weight model for coding and finance

  19. Moonshot AI launches Kimi K3.1 Pro open-weight model for coding and finance

  20. Moonshot AI launches Kimi K3.1 Pro, a 3-trillion-parameter open-weight model

Shared connections

Related stories