Alibaba 🇨🇳 · QVQ

QVQ-72B-Preview

An open-weight 72-billion-parameter preview model built on Qwen2-VL for extended multimodal reasoning in mathematics, science and complex visual problems.

DOOM SCORE63.6out of 100model risk profile, not the overall index
0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 3

Why this model scores 63.6

QVQ adds strong visual chain-of-thought reasoning over Qwen2-VL-72B but remains a preview with documented focus and hallucination limitations. Open weights raise misuse and control difficulty, while early distribution is lower than mature Qwen tiers.

Capability74
Autonomy24
Deployment78
Misuse potential69
Control difficulty72
MODEL-ATTRIBUTED EVIDENCE

News tied to QVQ-72B-Preview

The model score of 63.6 rates this model's risk profile. The overall Doom Index of 67.9 measures the complete temporally weighted evidence record. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.09

Each article's current Doom Index contribution is divided equally among the exact models named on that article. This prevents multi-model evidence from being claimed in full on several model pages. Model risk scores use a bounded temporal offset around their technical profile, but never feed back into the overall index.

AUDIT TRAIL

Model score history

  1. R3
    Doom Score 63.6

    Exact-version evidence chronology replayed after run doombench-hourly-news-20260812-142416 under temporal-monthly-pressure-v4.

    12 Aug 2026
  2. R2
    Doom Score 64.6

    Exact-version evidence chronology replayed after run doombench-hourly-news-20260812-132112 under fixed-sensitivity-v3.

    12 Aug 2026
  3. R1
    Doom Score 67.1

    Initial exact source-backed profile for QVQ-72B-Preview using the official December 25 date.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for QVQ-72B-Preview.
  1. QVQ-72B-Preview by Alibaba has a DoomBench model risk score of 63.6 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. QVQ-72B-Preview's highest current DoomBench dimension is deployment at 78.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to QVQ-72B-Preview, while keeping the model's risk profile separate from each item's contribution to the live Doom Index.

    https://www.doombench.com/models/alibaba-cloud-qvq-72b-preview