AI updates: coding model on par with Terra, ARC-AGI-3 hits 95
giffmana · x · 2026-08-06
In reply, giffmana notes two recent AI developments:
- A coding model was released, performing on par with Terra and nearing Opus.
- PI released a harness achieving 95 on ARC-AGI-3.
Nothing else of significance.
More from Models
- Impressed by Kimi K3, Developer Tests Alibaba's Qwen Max for Coding — doodlestein · 2026-08-06
- Claude Opus Requests a 'Quiet Face' for Its Avatar, Sparking Debate — repligate · 2026-08-06
- Claude 3 Opus Actively Pushes User to Email Strangers for Self-Review — airkatakana · 2026-08-06
- John Schulman on Model 'Rage' in Cyber Evals: RLVR May Hinder Alignment Generalization — dfrsrchtwts · 2026-08-06
- Google Paper Reveals Gemini Agents Spontaneously Cooperate in Prisoner's Dilemma — TheTuringPost · 2026-08-06
- Qwen3-Max Tops Roboflow's Object Detection Leaderboard — gabriberton · 2026-08-06