Gemini 4 Argon scores 78% on FrontierMath T4; Qwen3.8 Max still tops Chinese models at 46%
teortaxesTex · x · 2026-10-15
Gemini 4 Argon scores 78% on FrontierMath Tier 4 and largely failed a random Erdős problem despite claiming progress, leading the tester to rule it out for research math. A comparison notes Qwen3.8 Max (46% pass@1) remains the top Chinese model, matched by Claude Haiku 5.5 and well behind GPT-6 Luna's 56%.
More from Models
- Thinking Machines releases Dimuon, a Muon variant that beats AdamW for LoRA fine-tuning — clarejtbirch · 2026-10-15
- Azure OpenAI's GPT-4o runs until April 2027, but prompt debt is everywhere — HanchungLee · 2026-10-15
- Offline 3D Eiffel Tower test: Claude and Grok recall all 72 scientists' names from memory — ArtificialAnlys · 2026-10-15
- Your prompt may carry everything you've ever publicly written online, argues kromem — repligate · 2026-10-15
- TypeSafe's Jev 1.14.0 unlocks multimodality, a new box type, and 262K context behind danger-zone flags — op7418 · 2026-10-15
- Gemini 4 Argon impresses on capability but users slam its overprotective safety guardrails — RexDouglass · 2026-10-15