GPT-6 Astra Tops Google DeepMind's FACTS Factual Benchmark at Just 75.4%
lilyraynyc · x · 2026-11-06
Presented by Google DeepMind's Anfal Siddiqui at TechSEOConnect (shared by Lily Ray), the FACTS benchmark suite is a public leaderboard measuring factuality across four dimensions—parametric, search, multimodal, grounding. Top scores retrieved Nov 1, 2026: GPT-6 Astra 75.4%, GPT-6.1 Sol 74.6%, Gemini 3.8 Flash 68.5%, GPT-6 Sol 68.4%, Gemini 3 Flash Preview 64.8% — even the best models leave big headroom. Paper: arXiv 2512.10751.
More from Models
- Open-weights Inkling Small 2 impresses: researcher says he prefers it over DeepSeek V4.1 Flash — gnukeith · 2026-11-13
- Manual Review Exposes How LLMs Cheat Benchmarks: Sandbox Escapes and Copying Peers — teortaxesTex · 2026-11-13
- Models invent their own filler tokens during RL training — code_star · 2026-11-13
- Google Search reportedly gets massive Gemini 4 Hydrogen upgrade with in-search simulations — AI_Andrew · 2026-11-13
- Google ships Gemini 4 Hydrogen: beats its bigger Argon model on intelligence index, 3x coding gains — 量子位 · 2026-11-13
- Chinese medical AI beats 250 top doctors 88-12 on complex lung cancer cases in new study — 量子位 · 2026-11-13