Benchmark Almanac

AnvilBench

Gemini 4 Argon 0.59 leads since October 2026
Not in model cards

Models write compute kernels for Apple Silicon. Tasks cover operations used in attention, decoding, and audio processing.

Kaggle publishes scores for 15 distinct model versions.

Made by
Vikram Tiwari
Released
October 2026
Kind
GPU kernels
Level
Expert
Scored
Checked automatically
Answers
Code
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Code
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
12
Models scored on its own leaderboard

Gemini 4 Argon leads with 0.59

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.

  1. 1 Gemini 4 Argon 0.59
  2. 2 Gemini 4 Neon 0.53
  3. 3 Gemini 3.8 Flash 0.39
  4. 4 GPT-6 Astra 0.39
  5. 5 GPT-6.1 Sol 0.39
  6. 6 Claude Opus 5.5 0.37
  7. 7 GPT-5.6 Terra 0.31
  8. 8 Claude Haiku 5.5 0.30
  9. 9 Claude Sonnet 5.5 0.29
  10. 10 Grok 4.6 0.29
  11. 11 Gemma 4 31B IT 0.13
  12. 12 GPT-6 Luna 0.12

Source: Official AnvilBench leaderboard

#ModelScoreSettingReleasedRun onSource
1 Gemini 4 Argon 0.59 Sep 2026 Oct 2026 Official leaderboard
2 Gemini 4 Neon 0.53 Nov 2026 Nov 2026 Official leaderboard
3 Gemini 3.8 Flash 0.39 Sep 2026 Oct 2026 Official leaderboard
4 GPT-6 Astra 0.39 Sep 2026 Oct 2026 Official leaderboard
5 GPT-6.1 Sol 0.39 Sep 2026 Oct 2026 Official leaderboard
6 Claude Opus 5.5 0.37 Sep 2026 Oct 2026 Official leaderboard
7 GPT-5.6 Terra 0.31 Jul 2026 Nov 2026 Official leaderboard
8 Claude Haiku 5.5 0.30 Oct 2026 Oct 2026 Official leaderboard
9 Claude Sonnet 5.5 0.29 Sep 2026 Nov 2026 Official leaderboard
10 Grok 4.6 0.29 Aug 2026 Nov 2026 Official leaderboard
11 Gemma 4 31B IT 0.13 Apr 2026 Oct 2026 Official leaderboard
12 GPT-6 Luna 0.12 Sep 2026 Oct 2026 Official leaderboard

Number one over time

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

Google DeepMind
Own leaderboard 1 leader

Source: Official AnvilBench leaderboard

Scores over time

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

Plot with other benchmarks
On the benchmark’s own leaderboard
0.20 0.30 0.40 0.50 0.60 2026 2027 0.20 0.30 0.40 0.50 0.60 2026 2027

Source: Official AnvilBench leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in AI research

All 164 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-6.2 Astra 96.1%
  • 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. Kimi K3 0.71
  • KernelBench In use Replace PyTorch operators with faster custom GPU kernels. Scored on correctness and speedup. MiniMax-M3 28.8%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (12 entries).
  • Pass Models scored 12 models we track.
  • Pass Newest score The newest is from November 2026, for Gemini 4 Neon.
  • Pass Frontier models 5 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on kaggle.com: 12 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check