Moonshot AI releases Kimi K3, largest open-weight model at 2.8 trillion parameters
Source published ·Modelwire updated
Original coverage: Simon Willison ↗·How Modelwire adds context
The development
Moonshot AI's Kimi K3 marks a significant scaling milestone: at 2.8 trillion parameters, it becomes the largest open-weight model announced to date, surpassing DeepSeek's 1.6T offering. The model's self-reported benchmarks show competitive performance against frontier closed models like Claude Opus and GPT-5.5, though it trails the latest Claude Fable 5 and GPT-5.6. The July 27 open-weight release signals intensifying competition in the 3T-class tier, where Chinese labs are rapidly closing the capability gap with US incumbents. For practitioners, this represents both expanded inference options and a test case for whether scale alone sustains competitive advantage.
Modelwire’s AI-generated summary of coverage from Simon Willison.
Modelwire analysis
Analyst takeOur AI-generated reading of the wider context and the next developments to watch.
The benchmark framing buries the more consequential detail: Kimi K3's open-weight release on July 27 means the largest publicly available model in the world will be one that US frontier labs cannot easily match on openness, regardless of capability margins. That asymmetry matters more to enterprise buyers and fine-tuning shops than any leaderboard position.
This is largely disconnected from recent activity in our archive, so it belongs to a broader pattern worth naming directly. The 3T-class tier is becoming the new competitive front in the open-weight race, following DeepSeek's earlier 1.6T release establishing that Chinese labs are willing to publish at scales US labs have kept proprietary. Kimi K3 extends that logic one step further, and the July 27 date gives the market a concrete moment to pressure-test whether open-weight scale translates to real deployment adoption or remains a benchmark exercise.
If independent evaluators reproduce Kimi K3's self-reported scores on MMLU-Pro and GPQA Diamond within two weeks of the July 27 release, the capability claims hold. If those numbers slip materially on third-party runs, this is another case of labs optimizing for announcement benchmarks rather than general performance.
This interpretation is generated from the summary above and available source metadata. Our methodology · Report an error
MentionsMoonshot AI · Kimi K3 · DeepSeek · Claude Opus · GPT-5.5 · Claude Fable 5
How this coverage is produced
Modelwire uses AI to generate summaries and context from source headlines, snippets, and selected archive coverage. Automated checks do not verify every claim, and items are not routinely reviewed by a person before publication. Zacaria Solis operates the site. Read the linked source for the full evidence and report errors through our corrections process.
Modelwire summarizes, we don’t republish. Simon Willison originally reported this story as “Kimi K3, and what we can still learn from the pelican benchmark”. The full content lives on simonwillison.net. If you’re a publisher and want a different summarization policy for your work, see our takedown page.