Qwen3.8-27B scores 52, matches GPT-5.6 Luna and runs locally
udmrzn · x · 2026-08-19
Qwen3.8-27B scored 52 on the ArtificialAnalysis Intelligence Index, matching GPT-5.6 Luna (max). This open-weight model is strong enough to compete with the frontier and light enough to run locally, even in browsers with custom WebGPU kernels.
More from Models
- User preference: Sticking with Sonnet over Opus for coding — timfreemints · 2026-08-24
- Local DeepSeek V4 Flash Benchmark: 24 tok/s on Epyc + RTX 5090 — IntravenusDeMilo · 2026-08-24
- Speculation: Ox Alpha Allegedly a Multi-Party Collaboration — teortaxesTex · 2026-08-24
- Carnice-V3-27b: Beats 10x Larger Models Locally — TheMoonMidas · 2026-08-24
- Users Notice ChatGPT Becomes Less Agreeable, Stands Firm in Debates — USSGoat · 2026-08-24
- Grok ports DOOM to ESP32-S chip at high frame rate — yunta_tsai · 2026-08-24