MazeBench: Argon one-shots puzzles that stumped GPT-6 Astra, but walks away near the finish
patience_cave · x · 2026-10-15
MazeBench author shared details from the Gemini Argon 4 run: Argon one-shot difficult 2D sokoban puzzles that took GPT-6 Astra multiple attempts. Amusingly, Argon sometimes walks away when a puzzle is a few steps from solved — though it returns later to finish. The gem-collection chart shows GPT-6 Astra at 16%, Gemini 4 Argon 12%, with Opus 5.5 and GPT-6.1 Sol at 9%. Overall: agents have mastered the 2D portions but still struggle with 3D puzzles.
Related event: Gemini Argon 4 scores 12% on MazeBench, second only to GPT-6 Astra(3 posts)→
More from Models
- Opus 5 might be a very deep model or use looping, commentator speculates — scaling01 · 2026-10-16
- GPT-6.1 Sol audit finds 6% of arXiv papers had false main theorems, down to 1.5% recently — nitarshan · 2026-10-16
- Nvidia-backed post-training pushes Nemotron 3.5 VL past frontier on dermatology diagnosis — LoganGrasby · 2026-10-16
- Gemini 4 Argon scores 97.5% on ARC-AGI-3, second lab above 95% — patience_cave · 2026-10-16
- Microsoft MAI claims SOTA 3T-param in-house base models, post-training results coming — soldni · 2026-10-16
- Datadog's Scout 1.0: post-trained GLM-5.3 Flash matches frontier models for AI SRE at lower cost — Tim_Dettmers · 2026-10-16