LaunchAI ModelsJuly 20, 2026
Moonshot AI releases Kimi K3, a 2.8T parameter open-weights model

Kimi K3 is a 2.8-trillion-parameter open-weights model that rivals Claude Opus 4.8 on benchmarks like SpreadsheetBench 2. While it offers competitive performance, early reports indicate it is significantly slower and more token-intensive than frontier closed models.
10 sources
RT @ArtificialAnlys: Kimi K3 is second only to Fable 5 on AA-Briefcase, our agentic knowledge work...x.com
On Kimi K3: Its Capabilities And Related Discontentsthezvi.substack.com
Post-Kimi K3 and open weights models getting closer to the frontier again, I wonder if Anthropic and...bsky.app
Claude Fable 5 vs. Kimi K3: Same results, one-third the cost, 4x slowerthenewstack.io
‘Who’s Afraid of Chinese Models?’stratechery.com
Kimi K3 ranks #1 on @AfterQuery's SpreadsheetBench 2, surpassing Claude Fable 5reddit.com
Moonshot AI by email
Get an email when Moonshot AI ships something
More stories today
- Developer claims Claude Code now handles 95% of his work
- Semantica provides open-source enterprise intelligence layer for AI agents
- Best practices for creating professional-grade agent skills
- Bloomberg video explores Africa's role in AI infrastructure
- Creative Intelligence Suite provides agents for structured ideation