ModelsAug 18, 2026 · 03:534 sources

Alibaba Qwen3.8 27B Matches GPT 5.6 Luna Performance in First Local Model Frontier Score

From huggingnews.com

Image: huggingnews.com
AI;DR

The Alibaba Qwen3.8 27B open weight model achieved a score of 52 on the Artificial Analysis Intelligence Index, matching the performance of models like DeepSeek V4-Pro. The natively multimodal tool is licensed under Apache 2.0 and runs at 115 tokens per second on a single Nvidia RTX 5090, or 206.1 tokens per second using SGLang recipes.

Read the sourceSign in to suggest

Sources

+ 3 more
  1. SUPPORT@xenovacom“can run locally... even in your browser with custom WebGPU kernels!”x.com
  2. SUPPORT@sgl_project“206.1 tok/s decode on a single RTX 5090, with our NVFP4 plus DSpark”x.com
  3. SUPPORT@rohanpaul_ai“Qwen3.8-27B's official BF16 checkpoint is 55.6 GB”x.com
#alibaba#qwen#benchmark#open-source#multimodal#nvidiaMarkdown
Content history (1)
  • 2026-10-04 · Summary · vi · Wording fix
    Theo chỉ số Artificial Analysis Intelligence Index mới nhất, mô hình mã nguồn mở mới của …
    Theo chỉ số Artificial Analysis Intelligence Index mới nhất, mô hình mã nguồn mở mới của …

One ranked edition, every morning

The stories that matter today. Free, cancel anytime.

Telegram @aidr_today