Moonshot AI has released Kimi K3, a 2.8 trillion parameter model that claims to rival top-tier US models like Claude Opus 4.8 and GPT-5.5 in performance benchmarks. While it offers competitive pricing for its class, the release highlights the ongoing significance of specific evaluation metrics like the Pelican benchmark in assessing true reasoning capabilities.
Background
The release of Kimi K3 marks a significant milestone for Chinese AI labs in developing massive-scale open-weight models. It intensifies competition with US counterparts like Anthropic and OpenAI, particularly regarding cost-performance ratios and advanced reasoning evaluations.
- Source
- Simon Willison
- Published
- Jul 17, 2026 at 04:19 AM
- Score
- 7.0 / 10