Gemini 4 Scores Just 78% on FrontierMath Tier 4 as OpenAI Models Saturate It
haider1 · x · 2026-10-21
Gemini 4 (Argon high) scored only 78% (±6.5%) on FrontierMath Tier 4, far behind OpenAI models that have essentially saturated the benchmark: GPT-6.1 Sol (max) hit 100%, GPT-6 Astra variants 97.6%, and Claude Opus 5.5 (max) 95%. The author questions where Google's deep think is, given Google's strong research-level math progress months ago.
More from Models
- Nace.AI Launches Drex Max 2.0, a 27B Local Decision Model for Computer Use — DotaMate · 2026-10-21
- Speculation: Codex/Work also cap at 12 hours — OpenAI quietly eats cache-miss costs — dotey · 2026-10-21
- Using GPT-6 Pro for strategic business discovery: a 51-minute multi-subagent run and reusable prompt — cedric_chee · 2026-10-21
- ChatGPT's Live Video voice quietly switches you to GPT-4o, barely documented — RileyRalmuto · 2026-10-21
- QLoRA author: DeepSeek's open inference stack rivals OpenAI's and beats Anthropic's — DavidBennett__ · 2026-10-21
- Reflection's open-weight Beam 1.0 (501B-A23B) free on Nous Portal for a week — alexcovo_eth · 2026-10-21