DOOMBENCH.COM
OpenAI 🇺🇸 · GPT-6.1

GPT-6.1 Sol

GPT-6.1 Sol is a proprietary hosted general model released through the API, Codex and Work. OpenAI reports near-Astra agentic performance at lower prices and treats it as Critical for cyber and High for biological/chemical capability. Its exact system card documents safeguards plus residual persistence and monitoring limitations in controlled tests.

DOOM SCORE96.3out 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 96.3

Capability 99 reflects near-Astra coding, computer-use and professional-task results; autonomy 97 reflects stronger end-to-end tool and long-running workflow performance. Deployment 98 reflects broad hosted access and lower prices, tempered by proprietary access. Misuse 99 reflects Critical cyber and High biological/chemical classification despite layered safeguards. Control difficulty 88 reflects unwanted persistence and partial reasoning controllability, offset by stronger refusal, honeypot monitoring and lower severe-event rates in controlled evaluations. These comparative editorial dimensions are not probabilities.

Capability99
Autonomy97
Deployment98
Misuse potential99
Control difficulty88
MODEL-ATTRIBUTED EVIDENCE

News tied to GPT-6.1 Sol

The model score of 96.3 rates this model's risk profile. The overall Doom Index of 65.7 measures the complete temporally weighted evidence record. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.16

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

OpenAI releases GPT-6.1 Sol with Critical-rated cyber capability at lower cost

OpenAI released GPT-6.1 Sol to its API, Codex and Work on September 29. It reports near-Astra performance in agentic coding and computer use at one fifth Astra's standard API prices. Its system card treats cyber capability as Critical and biological/chemical capability as High, while reporting stronger safeguards and residual persistence risks in controlled evaluations.

Full item contribution
+0.16
GPT-6.1 Sol equal share
+0.16
Read assessment →
AUDIT TRAIL

Model score history

  1. R1
    Doom Score 96.3

    Create the exact publicly hosted GPT-6.1 Sol profile from its dated launch and system card, not an inferred revision of GPT-6 Sol.

    01 Oct 2026

Share this page

DoomBench social sharing card for GPT-6.1 Sol.

  1. GPT-6.1 Sol by OpenAI has a DoomBench model risk score of 96.3 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. GPT-6.1 Sol's highest current DoomBench dimension is capability at 99.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to GPT-6.1 Sol, 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-6-1-sol