Sol hits 84.5% reliability as Kimi peaks at 89.4% on the same benchmark
zainhas · x · 2026-07-23
A benchmark chart compares Sol and Kimi on a task set, showing a tradeoff between reliability and peak solve rate.
- Sol reaches a record reliability of 84.5% and solves 61 tasks 4/4, but has a lower peak coverage of 85.8%.
- Kimi reaches a higher peak pass@4 solve rate of 89.4%, but its reliability is lower at 76.6%.
- The chart frames them as opposite corners of the plane: one is steadier, the other has higher ceiling.
Related event: Kimi K3 Max vs GPT-5.6 Sol Max: A Routing Problem in Coding(10 posts)→
More from Models
- Benchmark scores drop from 89% to 19% on new evals — how benchmaxxing breaks leaderboard trust — airesearch12 · 2026-09-11
- 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