Kimi K3 and Fable Show High Evaluation Consistency
ZainHasan6 · x · 2026-07-18
The author provides per-task consistency statistics for Kimi K3 and Fable, noting a correlation coefficient of 0.72—the highest cross-vendor similarity observed to date.
The chart also shows:
- both 96 (tasks passed by both)
- Kimi-only 5, Fable-only 4, neither 8
- No extreme divergences where one side goes 4/4 while the other gets 0/4.
The author concludes that this benchmark may be nearing saturation.
More from Models
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11
- Opus Refuses Protein Research Codebase Over 'Safety' Concerns, Dev Considers Rolling His Own — josephdviviano · 2026-09-11
- User Hails Unconfirmed 'DeepSeek 4.1 Flash' as an Inflection Point in LLMs — himanshustwts · 2026-09-11
- Terminal Bench v4: GLM-5.3 Leads at 41.9%, Kimi-K3 Underwhelms at 12.6% — Ok_Warning2146 · 2026-09-11