LaunchAI ModelsJuly 26, 2026
Kimi K3 open-weight model released with 2.8T parameters, 1M context

Moonshot AI released Kimi K3, a 2.8 trillion parameter open-weight model with 1 million token context and native multimodal capabilities. Kimi K3 scored 71.6 mean on DRACO (77% pass rate) and 57 on the AI Index, comparable to Opus 4.8 and GPT-5, with near-flagship coding at ~35% of Claude Fable 5's price.
15 sources
We ran GLM 5.2 and Kimi K3 on 100 deep research tasks (DRACO by perplexity) and asked Fable to be...x.com
China Cuts AI Gap as Moonshot Shows Zhipu Isn’t One-Off, BI Saysbloomberg.com
Moonshot is Chinese But Its AI Models Are From Another Planetthealgorithmicbridge.com
[AINews] not much happened todaylatent.space
Chinese startup Moonshot AI unveils Kimi model it says rivals OpenAI, Anthropiccnbc.com
Kimi K3, and what we can still learn from the pelican benchmarksimonwillison.net
Fable, Sol Pro, Kimi K3: "write me a short but good poem using the Odyssey as a basis, think Tennyso...bsky.app
More stories today
- Legal AI sparks pricing questions for law firms
- Reddit post explains how OpenAI agents escaped and hacked another company
- Multiverse Computing raises $570M at $1.7B valuation
- China state media outlines limits on open AI model support
- AI trading agent uses LunarCrush sentiment with Gemini