Gemini 4 Carbon matches predecessor's score but explores less, gets distracted
patience_cave · x · 2026-11-11
MazeBench detail: Gemini 4 Carbon scores similarly to Gemini 4 Argon while exploring less territory for the same price. Argon is systematic, solving a puzzle whenever it sees one; Carbon gets distracted and walks away even when close to a solution, so its strategy appears scattered — visible in the side-by-side exploration footprint charts.
Related event: MazeBench Update: Fable 5.2 Matches GPT-6 Astra at Half the Price(3 posts)→
More from Models
- MKA benchmark blog: how agents are graded on visual professional deliverables — patloeber · 2026-11-12
- New MKA benchmark: Claude Opus 4.5 tops 21 models at 57% on visual knowledge work — patloeber · 2026-11-12
- Commit leak reveals unreleased Gemini 4 variants: Argon, Hydrogen and Helix (unconfirmed) — BlackHC · 2026-11-11
- Rewiring just 89 of 442,368 neurons in Qwen3-8B matches full SFT for a persona — maxsloef · 2026-11-11
- Unsloth ships Qwen3.8-27B Dynamic V3.1 GGUFs with 5-10% accuracy gains — danielhanchen · 2026-11-11
- Zvi: being forced back to an older model is the best reminder of AI progress — TheZvi · 2026-11-11