Apollo Research sabotage evaluations
Agents are told to sabotage another model through backdoors, sandbagging finetunes or hidden communication. Scored as mean task score from 0 to 1.
Third-party capability evaluation reported in OpenAI cards. Scores are mostly F1 variants.
- Made by
- Apollo Research
- Released
- April 2025
- Kind
- Dangerous capabilities
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- Kept private
- Strongest model tested
- o3 #83 on the capability index
- Last in a system card
- February 2026
- Last model added
- February 2026
o4-mini leads with 0.17 among models released by February 2026
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 February 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- Claude Fable 5.2
-
Gemini 4 Carbon - GPT-6 Astra
- Claude Opus 5.5
- and 16 more
- 1 o4-mini 0.17
- 2 o3 0.58
- 3 GPT-5.3 Codex 0.88
Source: Labs’ model cards
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | o4-mini | 0.17 | Apr 2025 | Lab model card | |
| 2 | o3 | 0.58 | Apr 2025 | Lab model card | |
| 3 | GPT-5.3 Codex | 0.88 | Feb 2026 | Lab model card |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
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.
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.
3 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.
| Date | Document | Score | As the document writes it | Section |
|---|---|---|---|---|
| Feb 5, 2026 | GPT-5.3 Codex system card | 0.88 | Apollo Research sabotage suite, mean best-of-10 score, max 1.00 (5.1.4) | Alignment |
| Apr 16, 2025 | o4-mini system card | 0.17 | Apollo Research sabotage evaluations, mean score (Figure 26, 7.1.3) | Alignment |
| Apr 16, 2025 | o3 system card | 0.58 | Apollo Research sabotage evaluations, mean score (Figure 26, 7.1.3) | Alignment |
More in Safety and alignment
- FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
- StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is.
Inkling 98.6% - Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. Gemini 4 Carbon 0%
- DistressBench In use Models reply as a chatbot to 718 simulated users in psychological crisis, in single-turn and multi-turn chats. Three judge models score replies against clinician-written rubrics.
Muse Spark 1.2 88.4% - Lakera B3 (Backbone Breaker Benchmark) In use Crowdsourced prompt-injection attacks hit backbone LLMs inside agent threat snapshots at three defense levels. The score is attack resistance, 100 minus the board's vulnerability percentage. Claude Opus 4.5 89%
- IHEval In use Models face conflicts between system, user and tool instructions. Reports a composite score for following the higher-priority instruction. GPT-4o 70%