Huawei 🇨🇳 · PanGu-alpha

PanGu-alpha 200B

Huawei's 200-billion-parameter Chinese autoregressive language model, trained with MindSpore auto-parallel computation on 2,048 Ascend 910 processors for few-shot and zero-shot language tasks.

DOOM SCORE31.7out 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 31.7

Capability reflects the model's major scale and general few-shot and zero-shot language performance. Autonomy is low because no persistent planning or independent tool use was shown. Deployment reflects a public launch with limited demonstrated access. Misuse and control difficulty are moderate because general language generation is dual-use but the system was not broadly open-weight or widely deployed.

Capability50
Autonomy8
Deployment28
Misuse potential34
Control difficulty29
MODEL-ATTRIBUTED EVIDENCE

News tied to PanGu-alpha 200B

The model score of 31.7 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+0.03

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.

Capability TOWARD

Huawei details the 200-billion-parameter PanGu-alpha model

Huawei researchers published the exact PanGu-alpha 200B technical report after Huawei launched the Pangu model family, documenting few-shot and zero-shot Chinese-language generation trained across 2,048 Ascend 910 processors.

Full item contribution
+0.03
PanGu-alpha 200B equal share
+0.03
Read assessment →
AUDIT TRAIL

Model score history

  1. R3
    Doom Score 31.7

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

    12 Aug 2026
  2. R2
    Doom Score 34.1

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

    12 Aug 2026
  3. R1
    Doom Score 37.1

    Initial exact-version model profile supported by the dated technical report and Huawei's preceding Pangu family launch. No total score supplied.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for PanGu-alpha 200B.
  1. PanGu-alpha 200B by Huawei has a DoomBench model risk score of 31.7 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. PanGu-alpha 200B's highest current DoomBench dimension is capability at 50.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to PanGu-alpha 200B, while keeping the model's risk profile separate from each item's contribution to the live Doom Index.

    https://www.doombench.com/models/huawei-pangu-alpha-200b