Dots3-note open-sourced: 280B-total/16B-active multimodal model scores perfect 42/42 on IMO 2026, 1/10 the size of rivals
xiaosun86 · x · 2026-08-14
dotsstudioai has open-sourced dots3-note preview, a multimodal model with 280B total parameters and 16B active, supporting 512K context, designed for real-world long-horizon agentic tasks. SGLang provides day-0 support.
The model powered rednote's IMO 2026 gold, scoring a perfect 42/42 across all six problems, a mark achieved by only 7 human contestants this year. It did so without human formalization (Lean) or proof assistants, using only reasoning and Python end-to-end.
It is among the first models to hit 42/42 under official grading, at roughly 1/10 the size of other models that achieved this (e.g., Kimi K3, GPT5.6 Sol, Fable 5).
Related event: Xiaohongshu Open-Sources dots3-note 280B Multimodal Model(14 posts)→
More from Models
- From botching 9.9 vs 9.11 to tackling the hardest math problems in two years — Yuchenj_UW · 2026-10-07
- OpenAI claims 372 unsolved problems cracked, averaging about 3 hours each — i_dg23 · 2026-10-07
- Benchmark: OpenAI Decisions API costs 2x more, 5-10% worse than Jev — xeophon · 2026-10-07
- Cagliostro V3.5 135M Dethrones SmolLM2-135M on Open SLM Leaderboard at 27.49 — Megneous · 2026-10-07
- "If you still think LLMs suck at math, it's a skill issue" — hot take gains traction — alejandroll10 · 2026-10-07
- Sauers_: Opus 5.5 shows fewer strong technical opinions than Fable 5.1 and 5 — Sauers_ · 2026-10-07