Comparing generations-old public models can't settle whether Anthropic or OpenAI leads
zacharynado · x · 2026-10-01
A snarky observation circulating on X: people argue endlessly about whether Anthropic or OpenAI is ahead, yet the benchmarks they cite compare models multiple generations behind each lab's latest internal versions — making the debate largely meaningless.
More from Fun
- Dev Suddenly Realizes OpenAI DevDay Granted Him a Big Pile of API Credits — JasonBotterill · 2026-10-01
- One Word Apart: Anthropic Tweaked Its Anti-Profanity System Prompt Between Opus 4.1 and 4.5 — steipete · 2026-10-01
- User gets Opus 5.5 to review Gemini 4 with its own conclusions — zsakib_ · 2026-10-01
- Canyon Overdrive: a Three.js-powered vaporwave canyon shooter you can play in the browser — AIandDesign · 2026-10-01
- Creator uses Opus 5.5 to co-direct a short vignette set in Tokyo — ebbyamir · 2026-10-01
- Researcher shares his favorite AI 'spell' now that anyone can cast them — generativist · 2026-10-01