agents benchmark
NanoGPT
NanoGPT is an OpenAI AI-self-improvement evaluation that measures whether models can optimize training recipes for small GPT-style models, part of the suite tracking progress toward accelerating internal research.
Updated Aug 7, 2026
Published models3
Registry coverage3
MetricScore
EvidenceB
NanoGPT leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | OP | 14.5% | 100.0% | 3 | C | |
| 2 | OP | 9.7% | 50.0% | 3 | C | |
| 3 | OP | 1.7% | 0.0% | 3 | C |
Score distribution
Top published rows on the benchmark's original scale.
NanoGPT highlights
The top published results on this benchmark's own scale.
What is NanoGPT?
Definition and scoring fields from the benchmark registry.
NanoGPT is an OpenAI AI-self-improvement evaluation that measures whether models can optimize training recipes for small GPT-style models, part of the suite tracking progress toward accelerating internal research.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- NanoGPT
- Modality
- text
- Primary category
- agents
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- nanogpt|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about NanoGPT.
Which model scores highest on NanoGPT?
GPT-5.6 Terra is currently ranked first with 14.5%.
What does NanoGPT measure?
NanoGPT is an OpenAI AI-self-improvement evaluation that measures whether models can optimize training recipes for small GPT-style models, part of the suite tracking progress toward accelerating internal research.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
3 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.