Chinese open-weight models near closed frontier on coding and math, reshaping enterprise choice
ingliguori · x · 2026-11-23
Chinese open-weight releases (DeepSeek, Kimi, Qwen) are lowering inference costs and closing in on leading closed models on coding, math, and some agentic tasks, while US labs still lead at the absolute frontier. Enterprise model selection now factors in architecture—self-host vs API—not just accuracy, per the accompanying "The Model Decision Gets Another Chair" infographic.
More from Models
- AI agents play a 3D murder mystery: GPT-6.2 solves it in 3m14s — venturetwins · 2026-11-23
- Users say GPT-5.6 was the last model 'with the sauce' as newer ones feel robotic — thederbiedone · 2026-11-23
- Pretraining perplexity alone can't rank base models anymore — stochasticchasm · 2026-11-23
- Keeping up with five flagship models: 'Opus 5.5, Fable 5.2, GPT-6.2 Astra... I'm tired' — Angaisb_ · 2026-11-23
- Commenter pushes back on perplexity claim: math gains likely came from RL, not pretraining — stochasticchasm · 2026-11-23
- Influencer's overall model ranking puts opus 5.5 on top, weighing cost and speed — MatthewBerman · 2026-11-23