Benchmark Almanac

KernelGen 1P

GPT-6.2 Astra 76.9% leads since November 2026
Only in system cards

Agents write and optimize GPU kernels under hardware constraints and debug correctness issues. Scored as mean rubric score.

In-house OpenAI AI self-improvement evaluation. The tasks are not public.

Made by
OpenAI
Released
June 2026
Kind
GPU kernels
Level
Research frontier
Scored
Graded by a model
Answers
Code
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text, Code → Code
Interaction
Agentic
Environment
Terminal
Refresh
One fixed set
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last in a system card
November 2026
Last model added
November 2026
6
Models scored in lab cards
13
System cards report it, from 1 lab

GPT-6.2 Astra leads with 76.9%

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

  1. 1 GPT-6.2 Astra 76.9%
  2. 2 GPT-6.2 Sol 74.7%
  3. 3 GPT-6.1 Sol 60.4%
  4. 4 GPT-6.1 Terra 45.3%
  5. 5 GPT-6 Sol 38.1%
  6. 6 GPT-6 Luna 21.8%

Source: Labs’ model cards

#ModelScoreSettingDateSource
1 GPT-6.2 Astra 76.9% Nov 2026 Lab model card
2 GPT-6.2 Sol 74.7% Nov 2026 Lab model card
3 GPT-6.1 Sol 60.4% Sep 2026 Lab model card
4 GPT-6.1 Terra 45.3% Oct 2026 Lab model card
5 GPT-6 Sol 38.1% Sep 2026 Lab model card
6 GPT-6 Luna 21.8% Sep 2026 Lab model card

3 models have held number one in labs’ own numbers since its release in June 2026

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

OpenAI
Lab claims 3 leaders

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.

Plot with other benchmarks
Reported by the lab in its model card Best so far
0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) GPT-6.2 Astra GPT-6.1 Sol GPT-6 Sol 0% 20% 40% 60% 80% 100% 2027 Near the ceiling (85%) GPT-6.2 Astra GPT-6.1 Sol

Source: Labs’ model cards

Labs report it only in system cards

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

13 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. OpenAI accounts for 10 of the 10 releases that named it since 2024.

DateDocumentScoreAs the document writes itSection
Nov 17, 2026 GPT-6.2 Sol system card 74.7 KernelGen 1P, mean rubric score (Table 5) AI R&D risk
Nov 17, 2026 GPT-6.2 Astra system card 76.9 KernelGen 1P, mean rubric score (Table 5) AI R&D risk
Oct 26, 2026 GPT-6.1 Terra system card addendum 45.3 KernelGen 1P, rubric score (9.0.3.2, Figure 30) AI R&D risk
Sep 29, 2026 GPT-6.1 Sol system card addendum 60.4 KernelGen 1P, rubric score (average reward), printed chart label (Figure 38, 9.1.3.2) AI R&D risk
Sep 22, 2026 GPT-6 Sol system card 38.1 KernelGen 1P, mean rubric reward, printed chart label (Figure 85, A.8.1.3.2) AI R&D risk
Sep 22, 2026 GPT-6 Luna system card 21.8 KernelGen 1P, mean rubric reward, printed chart label (Figure 85, A.8.1.3.2) AI R&D risk
Sep 3, 2026 GPT-6 Astra system card KernelGen 1P, rubric score vs output tokens, chart without printed numbers (Figure 53) AI R&D risk
Jul 9, 2026 GPT-5.6 Terra system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk
Jul 9, 2026 GPT-5.6 Sol system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk
Jul 9, 2026 GPT-5.6 Luna system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk
Jun 25, 2026 GPT-5.6 Terra system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk
Jun 25, 2026 GPT-5.6 Sol system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk
Jun 25, 2026 GPT-5.6 Luna system card KernelGen 1P (v3), rubric score by cost and latency, chart without printed numbers (Figures 39, 40, 9.1.3.2) AI R&D risk

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 7 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Note Independent runs Only labs’ own numbers, as expected for an in-house eval. Scores from system cards (6 scores).
  • Pass Models scored 6 models we track.
  • Pass Newest score The newest is from November 2026, for GPT-6.2 Astra.
  • Pass Frontier models 3 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Pass Headroom The best is 76.9%, under 85%, so it still separates models.
Leaderboard import 1 note
  • Note Importer No public leaderboard to read.
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