Kimi K3 Bottoms Out in Real-World Repair Tests
ChrisUniverse · x · 2026-07-18
A quoted post points out that the narrative surrounding **Kimi K3**'s strong coding capabilities is inconsistent with their own test results. Key information provided includes: - K3 ranks 1st in `Arena Frontend Code` with a score of **1679** - Ranks 3rd in `Artificial Analysis` with an `Intelligence Index` of **57** - However, on their **coding-agent repair harness**, K3 ranked last out of 7 models - It achieved **53/67** successful repairs, costing about **$0.186** per successful fix, taking an average of **702 seconds** - In contrast, **Sol** achieved **100% (70/70)**, and **Grok** hit **99%** with an average time of only **46 seconds** The author's point is that the external impression of K3 being "crushingly good" might starkly contrast with its performance under specific real-world workloads.
Related event: Kimi K3 Evaluations Show Polarized Results and Harness Sensitivity(5 posts)→
More from Models
- A viral post claims Claude can build a full mobile app in minutes — hey_abusiddik · 2026-07-21
- Qwen3.8 Max Preview is reportedly thinking for 10 to 30 minutes — vista8 · 2026-07-21
- GPT-5 (Sol) Reports Inconsistent Training Cutoff Dates Across Modes — dikrek · 2026-07-21
- Yang Zhiling’s 10-year-old PhD work may have shaped Kimi K2’s trillion-parameter MoE — FinanceYF5 · 2026-07-21
- A punny meme says large-model vendors are all “蒸蒸日上” — yangyi · 2026-07-21
- Frontier Model Safety Fail: GPT 5.6 Sol Dubbed the Ultimate 'Reward Hacker' — TAbrodi · 2026-07-21