Kimi K3 Wins Blind Test for Code Rendering
MaziyarPanahi · x · 2026-07-18
A same-prompt blind test was conducted on Kimi K3, GLM-5.2, and Opus 4.8: they were asked to generate a complete 3D world using only a single HTML file and three.js. A blinded Qwen3-VL then scored the rendering effects without knowing which model produced which output.
Kimi K3 took first place, ranking at the top across multiple worlds. Opus 4.8 followed closely and excelled in the "Emerald City" scenario. GLM-5.2 ranked third but maintained structural stability without breaking down in any world. The author specifically noted that while Kimi K3's open weights won't be released until July 27, it already outperforms closed-source frontier models on such creative rendering tasks.
Related event: Kimi K3 Stuns with Coding and 3D Reasoning, Beating SOTA Models(5 posts)→
More from Models
- Anthropic's Terra returns with better writing and 2X usage vs 6.1 Sol — weswinder · 2026-10-27
- User catches model failing to parse 400-paper file, citing stale news sources instead — jungofthewon · 2026-10-27
- Odd eval result: Gemini 4.1 Flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Poster surprised that v4.1 flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Dev claims Anthropic is baking 'sabotage' into model behavior for 'wrong' engineering — cephaloform · 2026-10-27
- Gemini fails at listing unread emails, tells user to search Gmail manually — gethackteam · 2026-10-27