Oct 22, 2026
That storage lets it reuse data for 95 of every 100 tokens and keep its costs low
Oct 19, 2026
Changes with no test run decreased a code benchmark result from 73.4 to 63.2.
Oct 14, 2026
Memory alone did not fix this. Prompt rules that steer the search did.
Oct 13, 2026
Thinking off still reasons in 99.69 percent of open answers. The fix is the question.
Oct 12, 2026
A NeurIPS 2026 paper from Tsinghua's thu-nics group explains why token-level routing has been stuck in notebooks, and ships a serving system that makes it 2–64× faster than what researchers had been running.
Oct 6, 2026
Chain-of-thought provides insight into AI reasoning. New research shows where that window starts to break.
Sep 29, 2026
Faster output, fewer tokens, and the API changes to check before switching.
Sep 24, 2026
Opus 5.5 and GPT-6 Sol, explained through five diagrams about agents, caching, and the cost of finished work.
Sep 22, 2026
Three regressions shrank by revision. Opus 5 left the coding charts. Biology fell by design.
Sep 18, 2026
Designing the system that assigns the work, checks the result, and knows when to stop.