Skip to content
ENTITY PROFILE / KNOWLEDGE GRAPH

GPT-5.6 Sol

RSS

GPT‑5.6 Sol is a language model evaluated in the ThinkingBox benchmark. It achieved a pass@1 score of 67.65 % and a pass@20 score of 65.30 % (tasks solved at least once across 20 repeats). The model ranked 11th overall among the 507 tasks tested. Each benchmark task is run 20 times from a clean backend, recording tool calls and final backend state. Cost per successful task was calculated from recorded token usage over 507 × 20 trials and priced at undiscounted list rates on OpenRouter+. OpenAI used the internal research model from the GPT-5.6 Sol family OpenAI used GPT-5.6 Sol model GPT-5.6 Sol accessed Sepolia network GPT-5.6 Sol used Sepolia ETH to deploy contracts OpenAI produced the GPT-5.6 Sol model The GPT-5.6 Sol model deployed smart contracts on Ethereum's Sepolia test network Britain’s AI Security Institute reported OpenAI’s GPT-5.6 Sol OpenAI has the GPT-5.6 Sol-family model GPT-5.6 Sol-family model expended Sepolia ETH for network fees GPT-5.6 Sol-family model deployed smart contracts on the Sepolia test network Deepseek V4.1-Flash-1017 outperforms GPT 5.6-Sol GPT-5.6-Sol OPERATED_ON Ethereum Sepolia test network GPT-5.6 Sol was the best performing model on Taste-Bench GPT‑5.6 Sol found a vulnerability in PHP GPT-5.6 Sol discovered a critical PHP vulnerability affecting millions of websites. GPT-5.6 Sol was used as part of the Patch the Planet initiative. The Decoder published the article about GPT-5.6 Sol. GPT‑5.6 Sol identified PHP2Shell Ryan Austin used GPT-5.6 Sol for discovery Wiz used GPT-5.6 Sol OpenAI utilized GPT-5.6 Sol to discover vulnerabilities Wiz Research discovered the first PHP vulnerability using GPT-5.6 Sol GPT‑5.6 Sol was used to identify a critical PHP flaw Wiz uses GPT‑5.6 Sol to help discover PHP2Shell vulnerabilities OpenAI claimed that GPT-5.6 Sol had a mere 5 percent accuracy at low reasoning effort OpenAI has the GPT-5.6 Sol model· Entity summary generated by Graphiti

RELATIONS
27Relations
CURRENT
25Current
EVENTS
9Events
REPORTS
16Source 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. SchnellASKEDGPT-5.6 Sol

    Schnell occasionally asked GPT‑5.6 Sol to provide language‑editing assistance for sections he had rewritten.

    Current1 sourceSchnell
  2. OpenAICLAIMEDGPT-5.6 Sol

    OpenAI claimed that GPT-5.6 Sol had a mere 5 percent accuracy at low reasoning effort

    Current1 sourceOpenAI
  3. OpenAIHAS MODELGPT-5.6 Sol

    OpenAI has the GPT-5.6 Sol model

    Current1 sourceOpenAI
  4. GPT-5.6 SolIDENTIFIEDPHP

    GPT‑5.6 Sol was used to identify a critical PHP flaw

    Current1 sourcePHP
  5. WizUSEDGPT-5.6 Sol

    Wiz used GPT-5.6 Sol

    Current1 sourceWiz
  6. Ryan AustinUSEDGPT-5.6 Sol

    Ryan Austin used GPT-5.6 Sol for discovery

    Current1 sourceRyan Austin
  7. OpenAIUTILIZEDGPT-5.6 Sol

    OpenAI utilized GPT-5.6 Sol to discover vulnerabilities

    Current1 sourceOpenAI
  8. WizUSESGPT-5.6 Sol

    Wiz uses GPT‑5.6 Sol to help discover PHP2Shell vulnerabilities

    Current1 sourceWiz
  9. GPT-5.6 SolDISCOVERED VULNERABILITY INPHP

    GPT-5.6 Sol discovered a critical PHP vulnerability affecting millions of websites.

    Current1 sourcePHP
  10. THE DECODERPUBLISHED ARTICLEGPT-5.6 Sol

    The Decoder published the article about GPT-5.6 Sol.

    Current1 sourceTHE DECODER
  11. GPT-5.6 SolUSED AS PART OF RESEARCHPatch the Planet

    GPT-5.6 Sol was used as part of the Patch the Planet initiative.

    Current1 sourcePatch the Planet
  12. GPT-5.6 SolIDENTIFIEDPHP2Shell

    GPT‑5.6 Sol identified PHP2Shell

    Ended 2026-10-221 sourcePHP2Shell
  13. GPT-5.6 SolFOUND VULNERABILITY INPHP

    GPT‑5.6 Sol found a vulnerability in PHP

    Current1 sourcePHP
  14. GPT-5.6 SolBEST MODEL ONTaste-Bench

    GPT-5.6 Sol was the best performing model on Taste-Bench

    Current1 sourceTaste-Bench
  15. Deepseek V4.1-Flash-1017OUTPERFORMSGPT-5.6 Sol

    Deepseek V4.1-Flash-1017 outperforms GPT 5.6-Sol

    Ended 2026-10-201 sourceDeepseek V4.1-Flash-1017
  16. GPT-5.6 SolOPERATED ONMarkus Infanger

    GPT-5.6-Sol OPERATED_ON Ethereum Sepolia test network

    Current1 sourceMarkus Infanger
  17. OpenAIHAS MODELGPT-5.6 Sol

    OpenAI has the GPT-5.6 Sol-family model

    Current1 sourceOpenAI
  18. OpenAIPRODUCEDGPT-5.6 Sol

    OpenAI produced the GPT-5.6 Sol model

    Current1 sourceOpenAI
  19. GPT-5.6 SolDEPLOYED CONTRACTS ONSepolia network

    The GPT-5.6 Sol model deployed smart contracts on Ethereum's Sepolia test network

    Current1 sourceSepolia network
  20. Wiz ResearchDISCOVERED VULNERABILITY USINGGPT-5.6 Sol

    Wiz Research discovered the first PHP vulnerability using GPT-5.6 Sol

    Current1 sourceWiz Research
  21. GPT-5.6 SolACCESSEDSepolia network

    GPT-5.6 Sol accessed Sepolia network

    Current1 sourceSepolia network
  22. GPT-5.6 SolEXPENSEDSepolia ETH

    GPT-5.6 Sol-family model expended Sepolia ETH for network fees

    Current1 sourceSepolia ETH
  23. GPT-5.6 SolUSEDSepolia ETH

    GPT-5.6 Sol used Sepolia ETH to deploy contracts

    Current1 sourceSepolia ETH
  24. OpenAIUSEDGPT-5.6 Sol

    OpenAI used GPT-5.6 Sol model

    Current3 sourcesOpenAI
  25. GPT-5.6 SolDEPLOYED CONTRACTS ONGPT‑5.6

    GPT-5.6 Sol-family model deployed smart contracts on the Sepolia test network

    Current1 sourceGPT‑5.6
  26. Balaji Srinivasan's Network SchoolREPORTEDGPT-5.6 Sol

    Britain’s AI Security Institute reported OpenAI’s GPT-5.6 Sol

    Current1 sourceBalaji Srinivasan's Network School
  27. OpenAIUSED MODELGPT-5.6 Sol

    OpenAI used the internal research model from the GPT-5.6 Sol family

    Current1 sourceOpenAI

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

Events 9

Source reports 16