GPQA Diamond — Epoch AI harness Effort: Low Configuration: Epoch AI harness 79.8% epoch.ai for OpenAI o3 (low) on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab) Archived copy (opens in a new tab) GPQA Diamond Accuracy Reported October 7, 2026 Evaluated July 15, 2026 at 3:15 AM UTC Benchmark leaderboard Epoch AI Independent Record for OpenAI o3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness GPQA Diamond — Epoch AI harness Effort: Medium Configuration: Epoch AI harness 80.8% epoch.ai for OpenAI o3 (medium) on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab) Archived copy (opens in a new tab) GPQA Diamond Accuracy Reported October 7, 2026 Evaluated August 7, 2026 at 12:42 AM UTC Benchmark leaderboard Epoch AI Independent Record for OpenAI o3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness GPQA Diamond — Epoch AI harness Effort: High Configuration: Epoch AI harness 81.8% epoch.ai for OpenAI o3 (high) on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab) Archived copy (opens in a new tab) GPQA Diamond Accuracy Reported October 7, 2026 Evaluated April 16, 2025 at 6:17 PM UTC Benchmark leaderboard Epoch AI Independent Record for OpenAI o3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness GPQA Diamond — Vals AI harness Effort: High Configuration: Vals AI harness: mean of zero-shot and few-shot CoT 84.1% vals.ai for OpenAI o3 (high) on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness (opens in a new tab) Archived copy (opens in a new tab) GPQA Diamond Accuracy Reported September 1, 2026 Benchmark leaderboard Vals AI Independent Record for OpenAI o3 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness HealthBench Effort: High 59.8% deploymentsafety.openai.com for OpenAI o3 (high) on HealthBench HealthBench (opens in a new tab) HealthBench Score Reported August 5, 2025 Model card OpenAI Self-reported Record for OpenAI o3 on HealthBench HealthBench HealthBench HealthBench Consensus Effort: High 92.8% deploymentsafety.openai.com for OpenAI o3 (high) on HealthBench HealthBench Consensus (opens in a new tab) HealthBench Score Reported August 5, 2025 Model card OpenAI Self-reported Record for OpenAI o3 on HealthBench HealthBench Consensus HealthBench HealthBench Hard Effort: High 31.6% deploymentsafety.openai.com for OpenAI o3 (high) on HealthBench HealthBench Hard (opens in a new tab) HealthBench Score Reported August 5, 2025 Model card OpenAI Self-reported Record for OpenAI o3 on HealthBench HealthBench Hard Humanity's Last Exam Full set — Scale AI evaluation Effort: Medium 19.2% labs.scale.com for OpenAI o3 (medium) on Humanity's Last Exam Full set — Scale AI evaluation (opens in a new tab) Archived copy (opens in a new tab) Humanity's Last Exam Accuracy Reported October 3, 2026 Benchmark leaderboard Scale AI Independent Record for OpenAI o3 on Humanity's Last Exam Full set — Scale AI evaluation Humanity's Last Exam Full set — Scale AI evaluation Effort: High 20.3% labs.scale.com for OpenAI o3 (high) on Humanity's Last Exam Full set — Scale AI evaluation (opens in a new tab) Archived copy (opens in a new tab) Humanity's Last Exam Accuracy Reported October 3, 2026 Benchmark leaderboard Scale AI Independent Record for OpenAI o3 on Humanity's Last Exam Full set — Scale AI evaluation Humanity's Last Exam Text only — Scale AI evaluation Effort: Medium 19.8% labs.scale.com for OpenAI o3 (medium) on Humanity's Last Exam Text only — Scale AI evaluation (opens in a new tab) Archived copy (opens in a new tab) Humanity's Last Exam Accuracy Reported October 7, 2026 Benchmark leaderboard Scale AI Independent Record for OpenAI o3 on Humanity's Last Exam Text only — Scale AI evaluation Humanity's Last Exam Text only — Scale AI evaluation Effort: High 20.6% labs.scale.com for OpenAI o3 (high) on Humanity's Last Exam Text only — Scale AI evaluation (opens in a new tab) Archived copy (opens in a new tab) Humanity's Last Exam Accuracy Reported October 7, 2026 Benchmark leaderboard Scale AI Independent Record for OpenAI o3 on Humanity's Last Exam Text only — Scale AI evaluation MMLU-Pro Original — Vals AI harness Effort: High Configuration: Vals AI harness 85.6% vals.ai for OpenAI o3 (high) on MMLU-Pro Original — Vals AI harness (opens in a new tab) Archived copy (opens in a new tab) Accuracy Reported September 1, 2026 Benchmark leaderboard Vals AI Independent Record for OpenAI o3 on MMLU-Pro Original — Vals AI harness MMMU-Pro Vals AI harness Effort: High Configuration: Vals AI harness 80.4% vals.ai for OpenAI o3 (high) on MMMU-Pro Vals AI harness (opens in a new tab) Archived copy (opens in a new tab) MMMU-Pro accuracy Setting: Standard (4 options) Reported September 1, 2026 Benchmark leaderboard Vals AI Independent Record for OpenAI o3 on MMMU-Pro Vals AI harness OSWorld Verified — 100 steps Configuration: 100-step limit 23.0% os-world.github.io for OpenAI o3 on OSWorld Verified — 100 steps (opens in a new tab) Archived copy (opens in a new tab) OSWorld-Verified success rate Reported July 28, 2025 Benchmark leaderboard OSWorld Independent XLang Lab Record for OpenAI o3 on OSWorld Verified — 100 steps OSWorld Verified — 15 steps Configuration: 15-step limit 9.1% os-world.github.io for OpenAI o3 on OSWorld Verified — 15 steps (opens in a new tab) Archived copy (opens in a new tab) OSWorld-Verified success rate Reported July 28, 2025 Benchmark leaderboard OSWorld Independent XLang Lab Record for OpenAI o3 on OSWorld Verified — 15 steps OSWorld Verified — 50 steps Configuration: 50-step limit 17.2% os-world.github.io for OpenAI o3 on OSWorld Verified — 50 steps (opens in a new tab) Archived copy (opens in a new tab) OSWorld-Verified success rate Reported July 28, 2025 Benchmark leaderboard OSWorld Independent XLang Lab Record for OpenAI o3 on OSWorld Verified — 50 steps SimpleQA Verified Original — Epoch AI harness Effort: High Configuration: Epoch AI harness 49.4% epoch.ai for OpenAI o3 (high) on SimpleQA Verified Original — Epoch AI harness (opens in a new tab) Archived copy (opens in a new tab) Accuracy Reported October 7, 2026 Evaluated August 27, 2026 at 7:28 PM UTC Benchmark leaderboard Epoch AI Independent Record for OpenAI o3 on SimpleQA Verified Original — Epoch AI harness SWE-bench Verified — Epoch AI harness Effort: Medium Configuration: Epoch AI harness 62.3% epoch.ai for OpenAI o3 (medium) on SWE-bench Verified — Epoch AI harness (opens in a new tab) Archived copy (opens in a new tab) Resolved Tasks Reported October 7, 2026 Evaluated February 12, 2026 at 3:46 AM UTC Benchmark leaderboard Epoch AI Independent Record for OpenAI o3 on SWE-bench Verified — Epoch AI harness SWE-bench Verified — mini-swe-agent Configuration: mini-swe-agent harness 58.4% github.com for OpenAI o3 on SWE-bench Verified — mini-swe-agent (opens in a new tab) Archived copy (opens in a new tab) Resolved Tasks Reported July 26, 2025 Benchmark repository SWE-bench Independent SWE-bench Team Record for OpenAI o3 on SWE-bench Verified — mini-swe-agent Toolathlon Original 17.0% toolathlon.xyz for OpenAI o3 on Toolathlon Original (opens in a new tab) Archived copy (opens in a new tab) Tasks completed Reported October 28, 2025 Benchmark leaderboard Toolathlon Independent Toolathlon Authors Record for OpenAI o3 on Toolathlon Original