TL;DR 1) ICPC: OpenAI 12/12, Gemini 10/12 (and one human‑unsolved) OpenAI says its general‑purpose reasoning models solved all 12 ICPC World Finals problems under the same five‑hour rules and judging as human teams; Gemini 2.5 Deep Think solved 10/12 and uniquely cracked one problem no human solved. Multiple reputable outlets and Google’s own blog corroborate these […] The post OpenAI edges Google at ICPC as xAI teases Grok 5 and OpenAI drops a bombshell “anti‑scheming” st...
