Microsoft 🇺🇸 · Phi-4

Phi-4

A 14-billion-parameter dense reasoning model with 16,000-token context, strong mathematical performance and Azure AI Foundry availability under a research release.

DOOM SCORE52.2out 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 4

Why this model scores 52.2

Phi-4 raises reasoning capability for a compact model and can support tool-using applications, but it is not independently agentic. Initial Foundry and research-license access limits deployment relative to later open releases, while small size raises potential diffusion.

Capability63
Autonomy16
Deployment70
Misuse potential55
Control difficulty56
MODEL-ATTRIBUTED EVIDENCE

News tied to Phi-4

The model score of 52.2 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+0.08

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

Microsoft introduces 14-billion-parameter Phi-4

Microsoft introduced Phi-4, a 14-billion-parameter reasoning model trained with synthetic and curated data and made available through Azure AI Foundry under a research release.

Full item contribution
+0.08
Phi-4 equal share
+0.08
Read assessment →
AUDIT TRAIL

Model score history

  1. R4
    Doom Score 52.2

    Source-backed model availability audit using the model's existing primary or authoritative catalogue evidence. Exact-version evidence chronology replayed under temporal-monthly-pressure-v4.

    12 Aug 2026
  2. R3
    Doom Score 52.2

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

    12 Aug 2026
  3. R2
    Doom Score 53.7

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

    12 Aug 2026
  4. R1
    Doom Score 57.1

    Initial exact source-backed profile for the December 12 Phi-4 release.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for Phi-4.
  1. Phi-4 by Microsoft has a DoomBench model risk score of 52.2 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. Phi-4's highest current DoomBench dimension is deployment at 70.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to Phi-4, while keeping the model's risk profile separate from each item's contribution to the live Doom Index.

    https://www.doombench.com/models/microsoft-phi-4