LIVE · refreshes every 20 min

updated Jul 22, 20:21 UTC

← 110010.art all stories

Hacker News (AI) ·1h ago

GPT-5.6 got smarter; then it kept acting.

Read the original at toloka.ai→

Original headline: "GPT-5.6 got smarter. Then it kept acting"

Coverage timeline

Jul 22, 18:58 UTC Hacker News (AI) lead source GPT-5.6 got smarter. Then it kept acting

Related stories

Drawing the Mona Lisa with GPT-5.6, Claude, Gemini, and Grok

Hacker News (AI) · 23h ago · source ↗

Moonshot AI unveils Kimi K3, a 2.8-trillion-parameter model described as their most capable to date, with open weights promised by July 27, 2026.

Simon Willison · 1d ago · source ↗

Relay-Bench evaluates LLMs on multi-domain reasoning chains; GPT-5.5 (xHigh) scores 43.3% on composite problems across domains

arXiv cs.CL · 16h ago · source ↗

We probed a pinned GPT-5.5 endpoint; every request carried about 1,447 hidden tokens

Hacker News (AI) · 4h ago · source ↗

Structured synthetic reasoning data improves arithmetic fine-tuning for small language models, using a 21,250-example GPT-5-mini-generated corpus derived from GSM8K

arXiv cs.AI · 16h ago · source ↗