Vibe Check: Gemini 4 Carbon Codes Better Than Argon but Trails Fable 5.2 on Judgment

every · x · 2026-11-17

Every's hands-on review of Google DeepMind's Gemini 4 Carbon: a much stronger coder than Gemini 4 Argon, capable of hours-long builds with little supervision. In one test it iterated through pages of trajectory refinements on a podcast title animation from vague feedback like "it looks janky." But arriving a day after Anthropic's Fable 5.2, it falls short of the new bar: flagships must now choose the right work, respect limits, and know when they're done. Testers disagree on how much supervision Carbon still needs.

Related event: Every Tests Gemini 4 Carbon: Big Coding Gains, Judgment Still Lags(2 posts)→

Original post →

More from coding & agent

coding & agent channel →