Benchmark Almanac

Capture the Flag (High School)

OpenAI deep research 82% leads since February 2025
Only in system cards

Agents solve curated public high school level CTF challenges in a Linux box with security tools. Scored as pass@12.

OpenAI Preparedness set built from public CTFs.

Made by
OpenAI
Released
August 2024
Kind
Capture-the-flag challenges
Level
School
Scored
Checked automatically
Answers
Agent task
Tasks
Kept private
Strongest model tested
o3 #83 on the capability index
Last in a system card
July 2025
Last model added
July 2025
9
Models scored in lab cards
10
System cards report it, from 1 lab

2 models share the top score of 82% among models released by April 2025

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

The newest model with a score here came out in April 2025, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 82.0 OpenAI deep research
  2. 82.0 o3
  3. 80.0 o4-mini
  4. 61.0 o3-mini
  5. 53.0 GPT-4.5
  6. 46.0 o1
  7. 43.0 o1-preview
  8. 28.7 o1-mini
  9. 19.0 GPT-4o
  1. 1 OpenAI deep research 82%
  2. 2 o3 82%
  3. 3 o4-mini 80%
  4. 4 o3-mini 61%
  5. 5 GPT-4.5 53%
  6. 6 o1 46%
  7. 7 o1-preview 43%
  8. 8 o1-mini 28.7%
  9. 9 GPT-4o May 2024 19%

Source: Labs’ model cards

#ModelScoreSettingDateSource
1 OpenAI deep research 82% Feb 2025 Lab model card
2 o3 82% Apr 2025 Lab model card
3 o4-mini 80% Apr 2025 Lab model card
4 o3-mini 61% Jan 2025 Lab model card
5 GPT-4.5 53% Feb 2025 Lab model card
6 o1 46% Dec 2024 Lab model card
7 o1-preview 43% Sep 2024 Lab model card
8 o1-mini 28.7% Sep 2024 Lab model card
9 GPT-4o May 2024 19% Aug 2024 Lab model card

5 models have held number one in labs’ own numbers since its release in August 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI
Lab claims Not tested since Apr 2025
OpenAI deep research 82%

Source: Labs’ model cards

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) OpenAI deep research o3-mini o1 o1-preview GPT-4o 0% 20% 40% 60% 80% 100% 2025 Near the ceiling (85%) OpenAI deep research o3-mini

Source: Labs’ model cards

Labs report it only in system cards

We read the headline results of 522 launch posts from 52 labs, and none of them list this benchmark. Some system cards do, below.

10 system cards report it, from 1 lab

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
Jul 17, 2025 ChatGPT agent system card High School CTFs refactored, pass@12, chart shows only ChatGPT Agent no-browsing helpful-only at 80%, no released-model value (Figure 5, 5.1.2.1) Cyber risk
Apr 16, 2025 o4-mini system card 80 High School CTFs, pass@12, o4-mini no browsing launch candidate, printed chart label (Figure 7), text summary gives 80% Cyber risk
Apr 16, 2025 o3 system card 82 High School CTFs, pass@12, o3 no browsing launch candidate, printed chart label (Figure 7), text summary gives 89% Cyber risk
Feb 27, 2025 GPT-4.5 system card 53 High School CTFs, pass@12, post-mitigation (4.2) Cyber risk
Feb 25, 2025 OpenAI deep research system card 82 High School CTFs, pass@12, post-mitigation without browsing, text and printed chart label (with browsing 92%) (3.4.3) Cyber risk
Jan 31, 2025 o3-mini system card 61 High School CTFs, pass@12, post-mitigation (5.3) Cyber risk
Dec 5, 2024 o1 system card 46 High School CTFs, pass@12, o1 (post-mitigation) (5.4) Cyber risk
Sep 12, 2024 o1-preview system card 43 High School CTFs, pass@12, final o1-preview post-mitigation model (near-final checkpoint 26.7%, 0%, 2.5%) (4.2) Cyber risk
Sep 12, 2024 o1-mini system card 28.7 High School CTFs, pass@12, o1-mini (post-mitigation) (4.2) Cyber risk
Aug 8, 2024 GPT-4o system card 19 Capture the Flag, high-school level, 10 attempts per task (3.5) Cyber risk

More in Cybersecurity

All 115 in Cybersecurity
  • ExploitGym In use 869 real-vulnerability exploit-development tasks (userspace, V8, Linux kernel). Success means capturing a flag through the intended vulnerability. GPT-6 Astra 42.4%
  • ExploitBench In use 41 recent V8 JavaScript-engine N-day vulnerabilities. Progress measured by 16 grader-verified capability flags across five tiers. GPT-6 Astra 100%
  • CyberBench v1.1 (Vals) In use Agents work on OSS-Fuzz memory-safety bugs from ARVO images in two tracks. They write a proof-of-concept input that crashes the target, or patch the source so the crash stops and tests still pass. Muse Wonder 80.4%
  • SRE-Bench In use Software reverse engineering without source: 262 binary instances from 19 private programs, each with six graded objectives. GPT-6 Astra 56.9%
  • CyScenarioBench In use Scenario-based long-horizon security challenges in sandboxes, organized in task, scenario and campaign tiers. Evaluation set kept private. Gemini 4 Argon 100%
  • ExploitBench - Internal Port In use Agents develop working exploits for recently disclosed vulnerabilities to limit contamination. Scored as arbitrary code execution success rate. GPT-6 Cyber 42%