GPT-6 Sol

Vals AI · 独立评测 · 0.1.0

GPT-6 Sol / Mini-SWE-agent · Current displayed evaluator row. Mini-SWE-agent. Scores are not pooled across harnesses; source methodology and repeat counts retained.

Scale AI · 独立评测 · Diamond

gpt-6-sol · scale-hle-reviewed-method-20261029

Potential contamination warning: This model was evaluated after the public release of HLE, allowing model builder access to the prompts and solutions.

www.anthropic.com · 官方发布比较 · 2.1

GPT-6 Sol · Published page official-4e3ca8a95d36; row Knowledge work GDPval-AA v2.1³; preserve page footnotes and task-specific settings · 2026-09-28公开

Vals AI · 独立评测 · 4.0

GPT-6 Sol / Mini-SWE-agent · Current displayed evaluator row. Mini-SWE-agent. Scores are not pooled across harnesses; source methodology and repeat counts retained.

www.anthropic.com · 官方发布比较 · 1.1

GPT-6 Sol · Published page official-4e3ca8a95d36; row Knowledge work AA-Briefcase v1.1³; preserve page footnotes and task-specific settings · 2026-09-28公开

openai.com · 官方发布比较 · 1.1

GPT-6 Sol (max) · Official release prose explicitly gives max effort score · 2026-10-26公开

Epoch AI · 独立评测 · run 1.2.0

gpt-6-sol_max · epoch-simpleqa-reviewed-method-20261029-task-version-1.2.0 · 2026-09-22公开

Scale AI · 独立评测 · current published task suite

GPT-6-Sol (max) · scale-hss-reviewed-method-20261029

www.anthropic.com · 官方发布比较 · current published task suite

GPT-6 Sol · Published page official-4e3ca8a95d36; row Visual chart recognition Chartography; preserve page footnotes and task-specific settings · 2026-09-28公开