OpenAI
Best on the Capabilities Index GPT-6 Astra 169.0
How does OpenAI stack up?
Every model on Epoch AI's Capabilities Index by release date, with OpenAI's in color.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)
OpenAI's best models
OpenAI Sep 2026
GPT-6 Astra
Closed weights Flagship
169.0
#3 of 256
Epoch Capabilities Index
Self-Reported Index 172.1 #3 of 404
- Humanity's Last Exam 54.8% #1 /47
- DeepSWE 74.1% #1 /28
- AutomationBench 41.4% #7 /42
OpenAI Sep 2026
GPT-6.1 Sol
Closed weights Mid-size
168.5
#5 of 256
Epoch Capabilities Index
Self-Reported Index 164.0 #15 of 404
- MultiChallenge 82% #1 /42
- Terminal-Bench 4.0 58.2% #5 /40
- GPQA Diamond 95.4% #6 /229
OpenAI Oct 2026
GPT-6.1 Terra
Closed weights
167.7
#6 of 256
Epoch Capabilities Index
Self-Reported Index 159.8 #32 of 404
- MultiChallenge 74.2% #7 /42
- Terminal-Bench 4.0 51.8% #8 /40
- GPQA Diamond 94.3% #11 /229
OpenAI Sep 2026
GPT-6 Sol
Closed weights Mid-size
163.8
#12 of 256
Epoch Capabilities Index
Self-Reported Index 161.9 #22 of 404
- AutomationBench 33.2% #8 /42
- Agents' Last Exam 32.2% #5 /31
- Terminal-Bench 4.0 49.4% #9 /40
OpenAI Jul 2026
GPT-5.6 Sol
Closed weights Flagship
163.0
#13 of 256
Epoch Capabilities Index
Self-Reported Index 162.6 #19 of 404
- MMMU-Pro 84.6% #1 /90
- DeepSWE 72.7% #4 /28
- AutomationBench 28.8% #11 /42
OpenAI Apr 2026
GPT-5.5 Pro
Closed weights
162.7
#14 of 256
Epoch Capabilities Index
- GPQA Diamond 93.9% #16 /229
- FrontierMath Tier 4 39.6% #2 /55
- ProtocolQA Open-Ended 36.7% #13 /27
What benchmarks does OpenAI mention most?
Here are the 28 benchmarks OpenAI mentions most in its launch posts.
MMLU GPQA Diamond MMMU MathVista AIME 2024 Codeforces rating SWE-bench Verified FrontierMath Humanity's Last Exam SWE-Lancer GraphWalks CharXiv Aider Polyglot MultiChallenge BrowseComp AIME 2025 MMMU-Pro tau2-bench Terminal-Bench 2.0 SWE-Bench Pro OpenAI MRCR v2 GDPval FrontierMath Tier 4 Toolathlon Agents' Last Exam DeepSWE OSWorld 2.0 AutomationBench
Source: Labs’ launch posts, read by Benchmark Almanac
Tacit Knowledge and Troubleshooting Capture the Flag (Professional) Standard Refusal Evaluation Capture the Flag (Collegiate) SWE-bench Verified StrongREJECT Long-form Biological Risk Questions SimpleQA Multimodal Troubleshooting Virology ProtocolQA Open-Ended Image input safety evaluations OpenAI PRs Cyber Range First-Person Fairness evaluation OpenAI Production Benchmarks and safe-completion evals TroubleshootingBench HealthBench HealthBench Hard HealthBench Consensus CVE-Bench OpenAI cyber safety evaluations (incl. Daybreak Blue) OpenAI Dynamic Mental Health Benchmarks Multiturn Jailbreak Evaluations HealthBench Professional OpenAI user-flagged hallucination eval Biology Model Refusal Evaluation ExploitBench SEC-bench Pro
Source: Labs’ system cards and technical reports, read by Benchmark Almanac
MMLU GPQA Diamond MMMU Tacit Knowledge and Troubleshooting SWE-bench Verified Capture the Flag (Professional) Standard Refusal Evaluation StrongREJECT SimpleQA ProtocolQA Open-Ended Multimodal Troubleshooting Virology Image input safety evaluations Humanity's Last Exam SWE-Lancer OpenAI Production Benchmarks and safe-completion evals TroubleshootingBench HealthBench HealthBench Hard HealthBench Consensus CVE-Bench OpenAI cyber safety evaluations (incl. Daybreak Blue) OpenAI Dynamic Mental Health Benchmarks Multiturn Jailbreak Evaluations HealthBench Professional OpenAI user-flagged hallucination eval ExploitBench SEC-bench Pro Biology Model Refusal Evaluation
Source: Labs’ launch posts, system cards and technical reports, read by Benchmark Almanac
Model cards
Every OpenAI model
- GPT-6 Cyber Nov 2026
- GPT-Image-2.6 Flare Nov 2026
- GPT-Image-2.6 Sunburst Nov 2026
- GPT-Rosalind-6 Oct 2026
- GPT-6.1 Terra Oct 2026
- GPT-6.1 Sol Sep 2026
- GPT-6 Luna Sep 2026
- GPT-6 Sol Sep 2026
- GPT-6 Astra Sep 2026
- GPT-6 Astra Pro Sep 2026
- GPT-Live-Transcribe Jul 2026
- GPT-5.6 Luna Jul 2026
- GPT-5.6 Sol Jul 2026
- GPT-5.6 Sol Pro Jul 2026
- GPT-5.6 Terra Jul 2026
- GPT-Live-1 Jul 2026
- GPT-Realtime-2.1 Jul 2026
- GPT-Realtime-2.1 Mini Jul 2026
- GPT-Realtime-2 May 2026
- GPT-5.5 Instant May 2026
- GPT-5.5 Apr 2026
- GPT-5.5 Pro Apr 2026
- GPT-5.4 Mini Mar 2026
- GPT-5.4 Nano Mar 2026
- GPT-5.4 Mar 2026
- GPT-5.4 Pro Mar 2026
- GPT-5.3 Instant Mar 2026
- GPT-Realtime-1.5 Feb 2026
- GPT-5.3-Codex-Spark Feb 2026
- GPT-5.3 Codex Feb 2026
- GPT-5.2 Codex Dec 2025
- GPT-5.2 Dec 2025
- GPT-5.2 Instant Dec 2025
- GPT-5.2 Pro Dec 2025
- GPT-5.1-Codex-Max Nov 2025
- GPT-5.1 Nov 2025
- GPT-5.1-Codex Nov 2025
- GPT-5.1-Codex-mini Nov 2025
- GPT-5.1 Instant Nov 2025
- GPT-5 Pro Oct 2025
- GPT-5-Codex Sep 2025
- gpt-realtime Aug 2025
- GPT-5 Aug 2025
- GPT-5 Chat Aug 2025
- GPT-5 mini Aug 2025
- GPT-5 nano Aug 2025
- gpt-oss-120b Aug 2025
- gpt-oss-20b Aug 2025
- ChatGPT agent Jul 2025
- o3-pro Jun 2025
- codex-1 May 2025
- o3 Apr 2025
- o4-mini Apr 2025
- GPT-4.1 Apr 2025
- GPT-4.1 mini Apr 2025
- GPT-4.1 nano Apr 2025
- GPT-4o Mar 2025 Mar 2025
- gpt-4o-mini-transcribe Mar 2025
- GPT-4o Mini TTS Mar 2025
- gpt-4o-transcribe Mar 2025
- o1-pro Mar 2025
- GPT-4.5 Feb 2025
- OpenAI deep research Feb 2025
- o3-mini Jan 2025
- GPT-4o Jan 2025 Jan 2025
- OpenAI Computer-Using Agent (CUA) Jan 2025
- o1 Dec 2024
- GPT-4o Nov 2024 Nov 2024
- o1-mini Sep 2024
- o1-preview Sep 2024
- GPT-4o Aug 2024 Aug 2024
- GPT-4o mini Jul 2024
- GPT-4o May 2024 May 2024
- GPT-4 Turbo Apr 2024 Apr 2024
- GPT-3.5 Turbo Jan 2024 Jan 2024
- GPT-4 Turbo (Jan 2024) Jan 2024
- Whisper large-v3 Nov 2023
- GPT-3.5 Turbo Nov 2023 Nov 2023
- GPT-4 Turbo Nov 2023 Nov 2023
- GPT-4V Nov 2023
- GPT-3.5 Turbo Instruct Sep 2023
- GPT-3.5 Turbo Jun 2023 Jun 2023
- GPT-4 Jun 2023 Jun 2023
- GPT-4 Mar 2023 Mar 2023
- GPT-3.5 Turbo (Mar 2023) Mar 2023
- davinci-003 Nov 2022
- davinci-002 Mar 2022
- InstructGPT 175B Jan 2022
- GPT-3 6.7B Jun 2020
- GPT-3 175B (davinci) May 2020
- GPT-2 (1.5B) Nov 2019
- DALL-E 3
- GPT-Image-1
- GPT-Image-1.5
- GPT-Image-1 mini
- GPT-Image-2
- GPT-Image-2.5 Flare
- GPT-Image-2.5 Sunburst
- Sora
- Sora 2
- Sora 2 Pro