.1 upgrade btw
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.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
. retweeted
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.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
very few things in life are as painful as having to live with the awareness of the cognitive limitations of your own brain, and by proxy the limitations of what things you can even do
. retweeted
AI is causing some immense tension for security as of late:
- OpenCode hacked lightning fast.
- All versions of PHP borked.
- LiteLLM library fucked again.
- eMudhra HTTPS certificates exposed as vulnerable.
- Node.js/Next.js/React/More with multiple high-severity CVEs.
- And so much more I am going insane.
what is it with valorant deathmatch players and holding angles all game instead of moving to try and simulate real duels
. retweeted
urodzony za późno na realne założenie rodziny i postawienie domu… ale za to urodzony w porę na ojebanie opakowania medikinetu i walenie konia 12h bez przerwy
idk what they did to claude code memory stuff, but its genuinely so much better than anything else ive ever used and it never gets in my way, unlike most of the other implementations
. retweeted
OpenCode 2.0
anoma.ly/notes/opencode-2-0/
