K3 may reach WeirdML top 10, with 75%–83% predicted across settings
teortaxesTex · x · 2026-07-28
The author asks for predictions on K3’s performance on WeirdML and says Harvard has a policy of not testing models served from China, leaving only two weeks to gauge its strength. Based on GLM’s trajectory, they would not be surprised if K3 enters the top 10, and they estimate a 75–83% result across settings.
Related event: K3 Projected to Hit WeirdML Top 10(2 posts)→
More from Models
- SkyPilot says serving Kimi K3 needs multi-node inference and a full stack — skypilot_org · 2026-07-28
- Fireworks says Kimi K3 matches Opus 5 quality at 2x–4.6x lower task cost — lqiao · 2026-07-28
- ARC-AGI-3 gains on Opus 5 may reflect harness improvements more than direct targeting — mhmazur · 2026-07-28
- Kimi K3’s 2.8T size makes the rumored 10T Fable claim look dubious — IndraVahan · 2026-07-28
- Kimi K3 tops Agent Arena among open-weight models, with zero tool hallucinations — arena · 2026-07-28
- Kimi K3 goes live on Together AI for long-running agentic workflows — togethercompute · 2026-07-28