AI · 1h ago
OpenAI harness tweaks triple ARC-AGI-3 scores; EU opens €10B AI gigafactory tender
OpenAI found that two harness settings—retained reasoning and context compaction—tripled GPT-5.6 Sol's ARC-AGI-3 score from 13.3% to 38.3%, solving all six levels. The EU opened a €10B tender for up to seven AI gigafactories, with bids due November 12. Cognition launched SWE-1.7 and other products in July.
Meridian48 take
The ARC-AGI-3 result underscores that benchmarks measure model-harness combinations, not just models, while the EU's gigafactory push signals a strategic shift toward compute sovereignty.
Read the full reporting
AI Daily Digest — August 1, 2026: ARC-AGI-3 Harness Discovery, EU AI Gigafactories, Devin SWE-1.7 →
DEV Community
arc-agi-3eu-ai-gigafactories