Moonshot AI Releases Kimi K2.6, Claims Open-Source Coding SOTA
Moonshot AI has released Kimi K2.6, an open-source coding model now live in chat mode and agent mode, available through its API, and published with weights and code. In its launch materials, the company said K2.6 achieves open-source state-of-the-art results on HLE with tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv with Python (86.7) and Math Vision with Python (93.2).
Moonshot is positioning K2.6 for long-horizon autonomous coding and agent workflows, saying it can handle more than 4,000 tool calls and over 12 hours of continuous execution, and support agent swarms with up to 300 parallel sub-agents and 4,000 steps per run. A launch example highlighted iterative local inference optimization on a Mac over 14 iterations, underscoring the model's focus on extended software and tool-use tasks.
From the sources (25 posts)
@scaling01Kimi-K2.6 is on HuggingFace
@scaling01Kimi-K2.6 Benchmarks
@scaling01Kimi-K2.6 SWE-Bench-Pro results are ridiculous
@testingcatalogMoonshot AI is rolling out Kimi K2.6 on Kimi Chat and APIs. All models got upgraded. - Kimi K2.6 Instant - Kimi K2.6 Thinking - Kimi K2.6 Agent - Kimi K2.6 Agent Swarm Did you get it already? 👀
@clementdelangueRT @scaling01: Kimi-K2.6 is on HuggingFace
@kimi_moonshotMeet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2) What's
@stevibeRT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench…
@scaling01Kimi-K2.6 Benchmarks (Blog is live:
@testingcatalog@Kimi_Moonshot Open-source SOTA? 👀👀👀 Testing time! 🔥
@scaling01Moonshot AI: "Our RL infra team used a K2.6-backed agent that operated autonomously for 5 days, managing monitoring, incident response, and system operations, demonstrating persistent context, multi-threaded task handling, and full-cycle e
@mweinbachRT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench…
@mweinbachKimi K2.6!!
@scaling01RT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench…
@zephyr_z9New Kimi drop
@testingcatalogKimi K2.6 from @Kimi_Moonshot is a new open-source SOTA on HLE with tools, SWE Bench Pro, and other benchmarks! - HLE w/ tools - 54.0 - SWE-Bench Pro - 58.6 - SWE-bench Multilingual - 76.7 Looks like it is testing time now 👀
@yuchenj_uwKimi K2.6 is open-source! Open-source SOTA: – SWE-Bench Pro: 58.6 – beats GPT-5.4 (xhigh) and Claude Opus 4.6 (max effort) I realize Kimi is shipping faster and faster. An S-tier open-source model team. Keep the open-source vs. closed-so
@andrewcurran_RT @Kimi_Moonshot: Meet Kimi K2.6: Advancing Open-Source Coding 🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench…
@kimmonismusHoly smokes, "state-of-the-art results on several coding and tool-use benchmarks" Open Source! Kimi is cooking! -54% HLE w tools -58.6% SWE Bench Pro -66.7% Terminal Bench tl;dr Moonshot AI released Kimi K2.6, an open-source model claimi
@huggingfacenew open coding sota 🤗
@clementdelangueRT @bridgemindai: Kimi K2.6 just dropped. And it crushed Claude Opus 4.6 on SWE-Bench Pro. Kimi K2.6: 58.6 GPT-5.4 xhigh: 57.7 Gemini 3.1…
@therundownaiAn open-source model just topped GPT-5.4, Claude Opus 4.6, and Gemini 3.1 Pro on some of the hardest benchmarks in AI. Moonshot AI just released Kimi K2.6. What it's good at: > Long-horizon coding (12+ hour autonomous runs) > Coordinating
@mattshumer_I'm not usually a fan of OSS models, but this new Kimi release looks pretty great (on the surface) and is priced extremely well. I plan to try it later today... how has it been so far for you?
@kimi_moonshotMeet Kimi K2.6 agent - Video hero section, WebGL shaders, real backends. From one prompt. 🔹 Video hero sections - cinematic aesthetic, auto-composited 🔹 WebGL shader animations - native GLSL / WGSL, liquid metal, caustics, raymarching 🔹
@kimi_moonshotVideo hero sections, built right in. K2.6 agent calls video generation APIs to create real cinematic footage for your hero, not stock placeholders. Composited into the page, synced to scroll, with shader overlays.
@kimi_moonshotSpeaks fluent WebGL shader. Writes GLSL / WGSL directly - fragment shaders, vertex shaders, noise, SDF, raymarching. Prompt: "a liquid-metal hero with soft caustics."