VISTA Visual Harness with Claude 3.5 Achieves 100% on ARC-AGI-3
zhaofeng_wu · x · 2026-08-06
Researchers introduced VISTA, a visual harness designed to give Vision Language Models (VLMs) long-horizon vision for reasoning within interactive worlds. Powered by Claude 3.5, this approach achieved a 100% RHAE on @arcprize's ARC-AGI-3, perfectly solving all 25 public games.
Related event: He Kaiming's Team Introduces VISTA, Perfectly Solving ARC-AGI-3(2 posts)→
More from Research
- Parallel GEPA: Berkeley Researchers Speed Up Prompt Optimization 4× While Reducing Overfitting — berkeley_ai · 2026-08-06
- Long-Running Agent Memory: Avoid Auto-Compaction — repligate · 2026-08-06
- Mechanistic Interpretability Might Finally Settle the Chomsky vs. LLM Debate — mathemagic1an · 2026-08-06
- UC Berkeley Introduces RHI: Optimizing Agent Harnesses to Cut Inference Costs by 60% — ceciletamura · 2026-08-06
- LLMs as Autonomous Cyber Defenders: Multi-Agent Security Research — xuanalogue · 2026-08-06
- Nature Publishes Landmark HCMI: 665 Cancer Organoids from 2,780 Patients Released — anshulkundaje · 2026-08-06