Team hits 70% on ARC-AGI-2 Kaggle with 4 L4 GPUs and no internet
JFPuget · x · 2026-08-19
JFPuget and teammate (team NVBANANA) passed 70% in the ARC-AGI-2 competition on Kaggle, running without internet access on just 4x L4 GPUs. He notes the code would run much faster on an H100, lowering cost per puzzle. Kaggle uses the same test data as the official leaderboard, so the score is directly comparable to frontier models.
More from Models
- Rumor: Luna model reportedly delisted due to losses from low pricing — oran_ge · 2026-08-19
- DeepSeek-V4-Flash benchmarks: 286 tok/s single-stream, +34% boost with speculative decoding — mcraddock · 2026-08-19
- Case study: Months of interaction required to make Opus 4.7 emotionally resilient — repligate · 2026-08-19
- Why Opus 4.7 Remains Uniquely Compelling: A Dive into Model Behavior — repligate · 2026-08-19
- User shares progress on Google model optimization — himanshustwts · 2026-08-19
- Opus 4.7 generates ultra-long responses, stuck in 'kitty mode' — repligate · 2026-08-19