Kimi K3 Chain of Thought Suspected of Distilling Claude
Sauers_ · x · 2026-07-17
Users discovered that Moonshot's Kimi K3 exhibits patterns strikingly similar to Anthropic's Claude when outputting its Chain of Thought (CoT), such as making formatted decisions at the end. This suggests that Kimi K3's training data might contain or be distilled from Claude's outputs.
Related event: Kimi K3’s “I’m Claude” replies spark alignment debate(7 posts)→
More from Models
- Google launches Gemini 3.5 Flash Cyber for CodeMender, with limited access for governments — GoogleAI · 2026-07-22
- Kimi K3 tops Gemini 3.6 Flash on four shared public benchmarks — ChrisGPT · 2026-07-22
- Google’s year-long pause in new base-model pretraining draws sharp criticism — teortaxesTex · 2026-07-22
- Current setup is 8,192 input tokens and 2,048 output tokens, with 8k/512 next — TheZachMueller · 2026-07-22
- Kimi K3 feels slower than K2.7, but stronger on long coding jobs and refactoring — Far-Presence2711 · 2026-07-22
- Poolside launches Laguna S 2.1 with 118B parameters and 8B active per token — Madisonkanna · 2026-07-22