OpenAI·released 2026-09-221 source
GPT-6 Sol benchmark scores: 17 benchmarks tracked, leading 1. 0% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 1 of 17 benchmarks·best result 100 on OTIS Mock AIME 2024-2025 (unverified)·0 of 17 independently reproduced
Head-to-headGPT-6 Sol vs GPT-6.1 SolLeads on: OTIS Mock AIME 2024-2025
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 100accuracy (%) | unverifiedT2 | 2026-09-22 |
| DTBench | 95.55accuracy (%) | unverifiedT1 | 2026-09-22https://conceptualreasoning.ai/dtbench |
| ARC-AGI | 95.5accuracy (%) | unverifiedT1 | 2026-09-22https://arcprize.org/leaderboard |
| GPQA diamond | 92.42accuracy (%) | unverifiedT2 | 2026-09-22 |
| FrontierMath-Tier-4-v2-Private | 90accuracy (%) | unverifiedT2 | 2026-09-22 |
| FrontierMath-Tiers-1-3-v2-Private | 89.82accuracy (%) | unverifiedT2 | 2026-09-22 |
| ARC-AGI-2 | 89.58accuracy (%) | unverifiedT2 | 2026-09-22 |
| ProofBench | 83accuracy (%) | unverifiedT1 | 2026-09-22https://www.vals.ai/benchmarks/proof_bench |
| LMCA | 69.51accuracy (%) | unverifiedT1 | 2026-09-22https://conceptualreasoning.ai/lmca |
| SimpleQA Verified | 60.7accuracy (%) | unverifiedT2 | 2026-09-22 |
| APEX-Agents | 54.3accuracy (%) | unverifiedT2 | 2026-09-22 |
| Chess Puzzles | 53.7accuracy (%) | unverifiedT2 | 2026-09-22 |
| EBR-bench | 53.33accuracy (%) | unverifiedT2 | 2026-09-22 |
| Mystery Game Puzzles | 51.53accuracy (%) | unverifiedT2 | 2026-09-22 |
| FrontierCode | 49.27accuracy (%) | unverifiedT1 | 2026-09-22https://cognition.com/frontiercode |
| Furniture Assembly | 40.48accuracy (%) | unverifiedT2 | 2026-09-22 |
| SpireClimb | 7.5accuracy (%) | unverifiedT2 | 2026-09-22 |
Claims drawn from cited facts, not live model generation.
Developed by OpenAI, this model is a 2026 release, placing its origin and vintage in that recent generation of OpenAI work.
2 cited facts
This model is scored on 17 tracked benchmarks and currently holds a top score on 1 of them. Its benchmark record is unverified: outside evaluation has neither confirmed nor disputed the reported results.
3 cited facts
On average, it trails the benchmark leader by 13.58 points under the disclosed harness, a wide gap that reads as well behind the leaders. It holds the top score on 1 benchmark, indicating genuine category leadership there but not broad front-of-pack standing.
2 cited facts
1 facts cross-checked across data sources: 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks
GPT-6 Sol is an AI model developed by OpenAI, released 2026-09-22. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
GPT-6 Sol has recorded scores on 17 benchmarks, each shown with its evidence status.
GPT-6 Sol has recorded scores on 17 benchmarks — GPQA diamond, OTIS Mock AIME 2024-2025, DTBench, LMCA, SimpleQA Verified, Chess Puzzles, and 11 more. The full table above shows each score with its evidence status.
0 of 17 recorded scores (0%) are independently reproduced rather than self-reported by the lab.
GPT-6 Sol has 17 tracked claims: 17 unverified.