No open models in top 25 anymore: MiMo #26, Qwen #28, GLM #30
teortaxesTex · x · 2026-10-02
Victor Taelin notes that no open models remain in the top 25 of ArtificialAnalysis rankings: the best is MiMo at #26, followed by Qwen (#28) and GLM (#30), and he suspects open-source models may never crack the top 10 again.
teortaxesTex adds that open models are increasingly conflated with Chinese models, that until this week there wasn't even a reference training stack for the top Chinese NPU, and that prior efforts were "hobbyism on comically small compute" — predicting 2027-28 could get interesting.
More from Models
- Fable 5.2 beats GPT-6 Astra: $7 in 2 min vs $98 in 24 min on same prompt — jradoff · 2026-11-16
- Qwen4 Max hands-on: matches GPT-6.1 Sol on agentic evals, trails on hard SWE at higher cost — haider1 · 2026-11-16
- Nous exec: avoid GPT 900k context variant — extra charges kick in past 272k — Teknium · 2026-11-16
- Rumor: Google's Gemini Argon and Carbon are both 12T-parameter models — oran_ge · 2026-11-16
- Qwen4 Max vs GPT-6.1 Sol at high reasoning effort: $0.0909 vs $0.132 — Angaisb_ · 2026-11-16
- FutureSim Q3: LLMs now genuinely improve from experience, Opus 5.5 tops at 50% — rickasaurus · 2026-11-16