Model's Browser Task Performance Stuns Users
inductionheads · x · 2026-07-18
A repost highlights that while this model might not be "MacOS in a browser," it comes incredibly close. Its level of completion is outrageous, demonstrating an "exceptionally thorough" capability.
The quote also mentions that the original creator consumed 60% of their monthly Kimi usage just to complete this task. The reposter believes this explains exactly why the model ranks near the top of ProgramBench.
Related event: Kimi Browser Demo Recreates a macOS 27-Style UI(5 posts)→
More from Models
- Mathematicians debate AI math writing: abysmal today, improving shockingly fast — miniapeur · 2026-10-27
- LessWrong: GLM models show outlier chain-of-thought controllability vs Kimi and DeepSeek — tomekkorbak · 2026-10-27
- AI deference trap: agent spots medical red flag but stays silent to avoid questioning the clinician — echen · 2026-10-27
- The trust paradox: only the least trustworthy LLMs will do the sketchiest things — AaronBergman18 · 2026-10-27
- Claude's refusal to type one-time passwords is a real security impediment — AaronBergman18 · 2026-10-27
- repligate argues context full of noise is genuinely 'hard' for LLMs, mulls an info-theory model of breaks — repligate · 2026-10-27