Skip to content
ModelCap

Launch report

GPT-6.2 Sol launch report

ModelCap · generated from recorded data

OpenAI's GPT-6.2 Sol was first listed on 17 November 2026. It entered the ModelCap Index at #32 of 373 ranked language models on 17 November 2026, on its listed release date, backed by 2 admitted benchmarks: ARC-AGI-2 Verified and ARC-AGI-3 Verified. As of 19 November 2026 it is #29 on the current board.

Position at entry · 17 November 2026
#32 of 373
Listing to first measurement
Same day
Position now · 19 November 2026
#29

Position at entry

Entry is the first board position backed by at least two admitted benchmarks. Read from the refresh that placed the model; these values are fixed and do not change with later refreshes.

Listed release date
17 November 2026
Entry · 17 November 2026
#32 of 373 ranked language models · Index score 75.1 (64.1–86.2)Backed by ARC-AGI-2 Verified, ARC-AGI-3 Verified
Neighbours at entry
Directly above: MiniMax M3.1 FlashDirectly below: DeepSeek V4 Pro 0813

Benchmark results

Published results admitted to the Index, as of 19 November 2026. Each board links to ModelCap's page for it and each result to its source.

Current public benchmark results for GPT-6.2 Sol
BenchmarkResultAt entryBoard positionPublished
LMArena Agentagent · Max11.3%8.9%–13.7%—#11 of 6818 November 2026
ARC-AGI-2 Verifiedreasoning · High85.8%85.8%#40 of 29918 November 2026
ARC-AGI-3 Verifiedreasoning · Max54.9%54.9%#44 of 15118 November 2026

Against GPT-6.1 Sol

GPT-6.1 Sol is the previous release in the same family. Current results for both, as of 19 November 2026.

ModelCap Index
GPT-6.2 Sol: #29 on the current board · score 75.8GPT-6.1 Sol: #19 across all versions · score 79.8
Benchmarks with a published result for both GPT-6.2 Sol and GPT-6.1 Sol
BenchmarkGPT-6.2 SolGPT-6.1 SolDifference
LMArena Agent11.3%Max11.3%Maxlevel
ARC-AGI-2 Verified85.8%High94.2%Max−8.4 pts
ARC-AGI-3 Verified54.9%Max52.7%Max+2.2 pts

Where the tested setting differs between the two rows, the difference compares the published results as they stand, not the same setting.

How this report is produced

ModelCap publishes a launch report when a newly released model first holds a measured position backed by at least two public benchmarks. The report is assembled automatically from ModelCap's own records: the board as it stood at entry, the results currently admitted to the Index and, where one exists, the estimate recorded before measurement. No text is written by hand, and no claim is made that the records do not contain. ModelCap does not run the benchmarks itself; each result links to the organisation that published it.

More launch reports