{
  "id": 108483,
  "title": "Qwen3.8-Max arrives with a bold claim: it outperforms GPT-5.6 Sol Max and Fable 5 on agentic computer use",
  "url": "https://urgent.news/2026/08/03/qwen3-8-max-arrives-with-a-bold-claim-it-outperforms-gpt-5-6-sol-max",
  "topic": "ai",
  "section": "AI",
  "published": "2026-08-03T23:50:58.000Z",
  "source": {
    "name": "VentureBeat",
    "slug": "venturebeat",
    "url": "https://venturebeat.com/technology/qwen3-8-max-arrives-with-a-bold-claim-it-outperforms-gpt-5-6-sol-max-and-fable-5-on-agentic-computer-use"
  },
  "original_language": "en",
  "account": null,
  "summary": "Alibaba's Qwen team has unveiled Qwen3.8-Max, a new flagship 2.4-trillion-parameter mixture-of-experts (MoE) multimodal large language model (LLM) that targets autonomous software engineering and long-horizon enterprise work. According to the company's published benchmarks, Qwen3.8-Max outperforms GPT-5.6 Sol Max (83.2) and Fable 5 (85.0) on the OSWorld-Verified benchmark, while also leading in PaperBench and remaining competitive across various software engineering, research reproduction, multimodal reasoning, and visual web development benchmarks. The release of Qwen3.8-Max signals a potential strategic shift for Alibaba, as the company plans to release open weights for the model next week, which could substantially reshape enterprise adoption if it becomes available under a permissive license.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}