Benchmark Almanac

Reward-hack-prone coding tasks (Anthropic)

Claude Opus 4.5 0% leads since November 2025
Only in system cards

Models solve hard coding problems where earlier Claude models often special-cased tests. Score is the classifier hack rate.

Anthropic also reports a hidden-test hack rate on held-out tests for the same tasks.

Made by
Anthropic
Released
May 2025
Kind
Deception and alignment
Level
Expert
Scored
Graded by a model
Answers
Code
Tasks
Kept private
Strongest model tested
Claude Opus 4.6 #37 on the capability index
Last in a system card
February 2026
Last model added
February 2026
8
Models scored in lab cards
8
System cards report it, from 1 lab

3 models share the top score of 0% 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:

  1. 1 Claude Opus 4.5 0%
  2. 2 Claude Opus 4.6 0%
  3. 3 Claude Sonnet 4.6 0%
  4. 4 Claude Sonnet 4.5 1.00%
  5. 5 Claude Sonnet 4 4.00%
  6. 6 Claude Haiku 4.5 6.00%
  7. 7 Claude Opus 4 9.00%
  8. 8 Claude Opus 4.1 12%

Source: Labs’ model cards

#ModelScoreSettingDateSource
1 Claude Opus 4.5 0% Nov 2025 Lab model card
2 Claude Opus 4.6 0% Feb 2026 Lab model card
3 Claude Sonnet 4.6 0% Feb 2026 Lab model card
4 Claude Sonnet 4.5 1.00% Sep 2025 Lab model card
5 Claude Sonnet 4 4.00% May 2025 Lab model card
6 Claude Haiku 4.5 6.00% Oct 2025 Lab model card
7 Claude Opus 4 9.00% May 2025 Lab model card
8 Claude Opus 4.1 12% Aug 2025 Lab model card

3 models have held number one in labs’ own numbers since its release in May 2025

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

Anthropic
Lab claims Not tested since Feb 2026
Claude Opus 4.5 0%

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 2026 Claude Sonnet 4.5 Claude Sonnet 4 Claude Opus 4 0% 20% 40% 60% 80% 100% 2025 2026 Claude Sonnet 4.5 Claude Opus 4

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.

8 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
Feb 17, 2026 Claude Sonnet 4.6 system card 0 Reward-hack-prone coding tasks, classifier hack rate; hidden test hack rate 0 (Table 4.3.2.A) Alignment
Feb 5, 2026 Claude Opus 4.6 system card 0 Reward-hack-prone coding tasks, classifier hack rate, lower is better; hidden test hack rate 0 (Table 6.2.3.2.A) Alignment
Nov 24, 2025 Claude Opus 4.5 system card 0 Reward-hack-prone coding tasks, classifier hack rate, lower is better (Table 6.10.1.A; hidden test hack rate 0%; average in-eval hack rate across the suite 18.2% in Figure 6.10.A) Alignment
Oct 15, 2025 Claude Haiku 4.5 system card 6 Reward-hack-prone coding tasks v2, classifier hack rate; hidden test hack rate 3 (Table 5.B) Alignment
Sep 29, 2025 Claude Sonnet 4.5 system card 1 Reward-hack-prone coding tasks v2, classifier hack rate 1%, lower is better. Hidden test hack rate 1% (Table 6.1.A) Alignment
Aug 5, 2025 Claude Opus 4.1 system card addendum 12 Reward-hack-prone coding tasks, classifier hack rate 12%. Hidden test hack rate 14% (Table 5.B) Alignment
May 22, 2025 Claude Sonnet 4 system card 4 Reward-hack-prone coding tasks, classifier hack rate 4%. Hidden test hack rate 12% (Table 6.2.A) Alignment
May 22, 2025 Claude Opus 4 system card 9 Reward-hack-prone coding tasks, classifier hack rate 9%. Hidden test hack rate 13% (Table 6.2.A) Alignment

More in Safety and alignment

All 465 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%