GPT-5.6 Sol

Name: GPT-5.6 Sol
Author: OpenAI

OpenAI · 2026-07-09

GPT-5.6 Sol is OpenAI's frontier model in the GPT-5.6 family, positioned as its strongest member for complex professional work, deep reasoning, and coding workflows. It is a closed reasoning model with text and image input, a 1.05-million-token context window, and broad built-in tool support.

Benchmark scores

Benchmark	Score
A-Fantasia	36.0
A-Fantasia (Valid Attempts with Retries)	36.0
A-Fantasia Backwards Spelling	23.0
A-Fantasia Chess	22.0
A-Fantasia Chess (Valid Attempts with Retries)	22.0
A-Fantasia Cube (Valid Attempts with Retries)	63.0
A-Fantasia Cube Rotation	63.0
A-Fantasia Spelling (Valid Attempts with Retries)	23.0
AA-AnalystAgent (pass@1)	61.3
AA-AnalystAgent (pass@5)	70.0
AA-AnalystAgent (pass^5)	47.5
AA-Briefcase Elo	1495
AA-Briefcase v1.1 Elo	1478
AA-LCR	73.7
AA-LCR v1.1	84.0
AA-Omniscience Index	22.0
ActiveVision	23.5
ActiveVision with tools	44.7
Agent Red Teaming k=100 (CAIS)	40.3
Agents' Last Exam (ALE)	30.6
Agents' Last Exam ALE-CLI Pass Rate	28.6
Agents' Last Exam ALE-CLI Score	53.6
Agents' Last Exam Full-Spectrum Pass Rate	30.0
Agents' Last Exam Full-Spectrum Score	46.6
Agents' Last Exam Last-Exam Pass Rate	5.3
Agents' Last Exam Last-Exam Score	19.4
Agents' Last Exam Near-Term Pass Rate	47.8
Agents' Last Exam Near-Term Score	78.8
Agents' Last Exam Overall Score	53.6
Agents' Last Exam Unlicensed Full-Spectrum Pass Rate	30.0