Skip to content
ENTITY PROFILE / KNOWLEDGE GRAPH

GPT-5.5

RSS

GPT-5.5 is one of the best overall models for DiG-bench The Decoder published the article about GPT‑5.5’s omission of negative results. The Decoder featured GPT‑5.5 in its multi‑model study. The article in The Decoder discusses GPT‑5.5’s performance in the study. The Decoder featured GPT‑5.5’s omission of negative results. GPT‑5.5 and Opus 4.8 were evaluated in the same study on AI model honesty. GPT‑5.5 detects errors in every single case in the study. GPT-5.5 was the second best performing model on Taste-Bench GPT‑5.5 is used as the customer‑simulator model in τ²‑bench tasks MAI-DxO achieved best results with OpenAI's GPT‑5.5 OpenAI publishes the GPT-5.5 model, part of the Nemotron 3.5 Super VL family. Managed Deep Agents v0.10’s subagent definition uses the GPT-5.5 model from OpenAI.· Entity summary generated by Graphiti

RELATIONS
12Relations
CURRENT
10Current
EVENTS
5Events
REPORTS
6Source reports
KNOWLEDGE ATLAS0 entities · 0 relations
No public neighbors yet
Explore all published links of a node; previously discovered relations remain visible.

Relation timeline

BY VALID-FROM DATE
  1. GPT-5.5PUBLISHED BYOpenAI

    OpenAI publishes the GPT-5.5 model, part of the Nemotron 3.5 Super VL family.

    Current170 sourcesOpenAI
  2. Managed Deep AgentsUSES MODELGPT-5.5

    Managed Deep Agents v0.10’s subagent definition uses the GPT-5.5 model from OpenAI.

    Current64 sourcesManaged Deep Agents
  3. MAI-DxOBEST RESULTS FROMGPT-5.5

    MAI-DxO achieved best results with OpenAI's GPT‑5.5

    Current1 sourceMAI-DxO
  4. GPT-5.5USED AS CUSTOMER SIMULATOR INτ²‑bench

    GPT‑5.5 is used as the customer‑simulator model in τ²‑bench tasks

    Current1 sourceτ²‑bench
  5. GPT-5.5SECOND BEST MODEL ONTaste-Bench

    GPT-5.5 was the second best performing model on Taste-Bench

    Ended 2026-10-231 sourceTaste-Bench
  6. GPT-5.5PUBLISHED ARTICLE CONCERNINGTHE DECODER

    The article in The Decoder discusses GPT‑5.5’s performance in the study.

    Current1 sourceTHE DECODER
  7. GPT-5.5FEATURED INTHE DECODER

    The Decoder featured GPT‑5.5 in its multi‑model study.

    Current1 sourceTHE DECODER
  8. GPT-5.5EVALUATED IN SAME STUDYOpus 4.8

    GPT‑5.5 and Opus 4.8 were evaluated in the same study on AI model honesty.

    Current1 sourceOpus 4.8
  9. GPT-5.5PUBLISHED ARTICLE CONCERNINGTHE DECODER

    The Decoder published the article about GPT‑5.5’s omission of negative results.

    Current1 sourceTHE DECODER
  10. GPT-5.5FEATURED INTHE DECODER

    The Decoder featured GPT‑5.5’s omission of negative results.

    Current1 sourceTHE DECODER
  11. GPT-5.5DETECTED ERRORS IN EVERY CASEOpus 4.8

    GPT‑5.5 detects errors in every single case in the study.

    Current1 sourceOpus 4.8
  12. GPT-5.5BEST MODEL INxAI

    GPT-5.5 is one of the best overall models for DiG-bench

    Ended 2026-10-231 sourcexAI

Relations are facts Graphiti extracted from reports, not verbatim quotes. An end date means later reporting showed the relation no longer holds.

Events 5

Source reports 6