DOOMBENCH.COM
OpenAI 🇺🇸 · GPT-Image-2.5

GPT-Image-2.5 Sunburst

GPT-Image-2.5 Sunburst is OpenAI's higher-quality hosted image model, released through the API and the Images 2.5 product family. It improves editing consistency, reference fidelity, infographics, text rendering, and photorealistic output while using a layered safety stack.

DOOM SCORE64.2out 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.

Send detailed or private feedback →
CURRENT ASSESSMENT · REVISION 1

Why this model scores 64.2

Sunburst is the higher-capability Images 2.5 variant and can produce more convincing realistic synthetic media, which raises misuse potential despite low independent autonomy. OpenAI documents multiple prompt, input, and output safety layers, provenance metadata, invisible watermarking, and hosted API access. The stronger output capability warrants a somewhat higher misuse and control-difficulty assessment than Flare while remaining more controllable than open weights.

Capability86
Autonomy4
Deployment90
Misuse potential82
Control difficulty59
MODEL-ATTRIBUTED EVIDENCE

News tied to GPT-Image-2.5 Sunburst

The model score of 64.2 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+0.06

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.

Misuse TOWARD

OpenAI releases more realistic Images 2.5 models with deepfake safeguards

OpenAI released GPT-Image-2.5 Flare and Sunburst across ChatGPT, ChatGPT Work, Codex, and the API. The system card says greater realism can enable more convincing political, sexual, and sensitive deepfakes, while layered filters, C2PA metadata, and invisible watermarking mitigate misuse.

Full item contribution
+0.11
GPT-Image-2.5 Sunburst equal share
+0.06
Read assessment →
AUDIT TRAIL

Model score history

  1. R1
    Doom Score 64.2

    Adds OpenAI's exact higher-quality Images 2.5 API model with source-backed capability, access, misuse, and control details.

    10 Sept 2026

Share this page

DoomBench social sharing card for GPT-Image-2.5 Sunburst.

  1. GPT-Image-2.5 Sunburst by OpenAI has a DoomBench model risk score of 64.2 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. GPT-Image-2.5 Sunburst's highest current DoomBench dimension is deployment at 90.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to GPT-Image-2.5 Sunburst, while keeping the model's risk profile separate from each item's contribution to the live Doom Index.

    https://www.doombench.com/models/openai-gpt-image-2-5-sunburst