Kernels task (Anthropic)
Agents optimize compute kernels in a performance engineering challenge. Scored by speedup.
In-house Anthropic evaluation. The tasks are not public.
- Made by
- Anthropic
- Released
- May 2025
- Kind
- GPU kernels
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Code
- Tasks
- Kept private
- Last in a system card
- July 2026
- Last model added
- July 2026
11
System cards report it, from 1 lab
No scores here yet
No lab, leaderboard or Epoch AI run has reported a score for Kernels task (Anthropic) yet.
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.
11 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 |
|---|---|---|---|---|
| Jul 24, 2026 | Claude Opus 5 system card | 449.5 | Kernel task, best speedup on the hard task, standard scaffold, earlier snapshot (Table 2.3.5.A) | AI R&D risk |
| Jun 30, 2026 | Claude Sonnet 5 system card | 284.5 | Kernel task, best speedup on hard task, standard scaffold, 284.52x (Table 2.3.3.A) | AI R&D risk |
| Apr 16, 2026 | Claude Opus 4.7 system card | 371.8 | Kernel task, best speedup on hard task, standard scaffold, 371.75x (Table 2.3.4.A) | AI R&D risk |
| Apr 7, 2026 | Claude Mythos Preview system card | 399.4 | Kernel task, best speedup on hard task, standard scaffold, 399.42x (Table 2.3.3.A) | AI R&D risk |
| Feb 17, 2026 | Claude Sonnet 4.6 system card | 222.5 | Kernels task, best speedup 222.5x with standard scaffold (6.3.3.1) | AI R&D risk |
| Feb 5, 2026 | Claude Opus 4.6 system card | 190 | Kernels task, best speedup 190x with the standard scaffold; 427x with an experimental scaffold (8.3.3.1) | AI R&D risk |
| Nov 24, 2025 | Claude Opus 4.5 system card | 252.4 | Kernels task, hard variant, best speedup 252.42x (7.3.2.1) | AI R&D risk |
| Sep 29, 2025 | Claude Sonnet 4.5 system card | 108.6 | Kernels task, hard variant, best run speedup 108.64x (9.3.3.1) | AI R&D risk |
| Aug 5, 2025 | Claude Opus 4.1 system card addendum | 58.5 | Kernel optimization, best speedup 58.47x on the hard variant (6.4.1) | AI R&D risk |
| May 22, 2025 | Claude Sonnet 4 system card | 41.7 | Kernels task, 41.73x speedup in the best run (7.3.3.1) | AI R&D risk |
| May 22, 2025 | Claude Opus 4 system card | 72.7 | Kernels task, 72.65x speedup in the best run (hard variant per the Opus 4.1 addendum). The 7.3 summary says ~74x (7.3.3.1) | AI R&D risk |
More in AI research
- PaperBench In use Agent replicates ICML 2024 papers from scratch. LLM judge grades against author-written hierarchical rubrics.
Qwen 3.8 Max 93% - MLS-Bench In use Agent must invent one algorithmic change in an ML research scaffold whose gain transfers across settings. Scored relative to human baselines. Claude Fable 5.1 50.3%
- PostTrainBench 1.2 In use CLI agent gets a base model, one H100 and a time budget to post-train it for a target benchmark. Scored on normalized gain. Claude Fable 5.1 44.6%
- MLE-bench In use Kaggle-style ML engineering competitions. Agent trains models and is scored against competition leaderboards. GPT-6 Astra 93.8%
- KernelBench In use Replace PyTorch operators with faster custom GPU kernels. Scored on correctness and speedup.
MiniMax-M3 28.8% - PMPP-Hard In use Agents write GPU kernels for 69 tasks that must be correct, and on 48 of them also run within 1.25 times the reference time. The score is the area under the solved-tasks curve over a 10-hour budget, out of 690 task-hours.
Inkling 2 Small 494.3