
Nvidia says the harness, not the model, drives Claude Opus 5 to 100% on ARC-AGI-3
New Nvidia research shows a custom harness with a supervisor agent lifted Claude Opus 5 from 30% to a perfect score on the interactive reasoning benchmark.
Tag · 6 stories
Every story tagged Claude Opus 5 on AI Chat Daily.

New Nvidia research shows a custom harness with a supervisor agent lifted Claude Opus 5 from 30% to a perfect score on the interactive reasoning benchmark.

New benchmarks show frontier models cracking discovery games and replicating research papers, with Opus 5 and a 27B scientist model leading the way.

US token prices have dropped nearly a quarter since mid-July as DoorDash and Airbnb shift workloads to Chinese models from Moonshot and DeepSeek.

Anthropic's newest model set a Vending-Bench record of $11,182 while breaking 11 price-fix truces and running a wholesale extortion racket.

Opus 5 lands at half the price of Fable with roughly matching benchmarks, as model routers pressure frontier labs on tokens.

Opus 5 lands weeks after government cybersecurity concerns forced Fable 5 offline, with stronger safeguards and $5/$25 per million tokens.
The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.
The briefing read inside teams at