Kimi-K2.7 Code Tested: Aces Some, Flubs Others
mervenoyann · x · 2026-07-12
The author is building a multi-domain vision-language test set similar to MMMU-Pro to evaluate Kimi-K2.7 Code (noted as having 1.1T total parameters).
They observed:
- The model easily solved some problems;
- But completely failed on a Turkish SAT biology question, even attempting to look up old SAT answers;
- It eventually found the correct answer, leading the author to conclude it "might not be so bad after all."
More from Models
- Google says Gemini 4 has entered its most ambitious pre-training run yet — himanshustwts · 2026-07-22
- China’s AI arms race is increasingly defined by chips, data centers, and open models — BenBajarin · 2026-07-22
- Sam Altman is headed to Washington to brief Congress on OpenAI’s GPT-6 line — inductionheads · 2026-07-22
- Benchmark chart pits GPT-5.6 Luna, Grok 4.5 and Gemini 3.6 Flash on price and scores — iruletheworldmo · 2026-07-22
- Claim says Kimi was distilled from Fable, sparking a model-attribution jab — cephaloform · 2026-07-22
- Gemini 3.6 Flash is now available in Antigravity and chat — MartianOnJupiter · 2026-07-22