Benchmark score mix-up: 74.2 was the old model's result, not the new one
dejavucoder · x · 2026-11-08
A light drama moment in the AI eval scene: someone cited a 74.2 benchmark score as a new result, only to be corrected that 74.2 was the old model's score. dejavucoder joins in with a quip suggesting they just rerun the benchmark themselves.
More from Fun
- AI-generated Lost Star Wars Audition Tapes parody nails 80s nostalgia — TinfoilTricorn · 2026-11-08
- Four years ago ChatGPT couldn't solve a text maze — a stark look at how far models have come — desdae · 2026-11-08
- signull: Philosophy is an underrated source of product insight, as Anthropic opens agent memory post with Kierkegaard — signulll · 2026-11-08
- 'We'll ban AI, and use AI surveillance to check if you're using the AI we banned' — ChrisGPT · 2026-11-08
- Dev vents about batch-generating 100 AI pages and prompting 100 times to deslop them — yuwen_lu_ · 2026-11-08
- Dev tells ChatGPT's Dot to stop being proactive — it nearly sent a WoW leveling guide — Dimillian · 2026-11-08