LLM Stats (NanoGPT): leaderboard

OpenAI's internal NanoGPT evaluation, as tracked by LLM Stats: agents modify code and tune hyperparameters to reduce the time needed to train a small language model from scratch to a target validation quality under fixed compute and time limits. Scored in percent as mean share of baseline training time saved (higher is better). LLM Stats combines GPT-6.2 Astra and Sol scores from the GPT-6.2 system card (Table 5, version 5) with GPT-5.6 Terra, Sol and Luna scores from the GPT-5.6 launch post; the system card notes evaluation versions differ slightly across releases.

Metric: Score (%). Source: llm-stats.com. Saturation forecast: Around November 2027. 5 models tracked. Updated 19 Nov 2026.

Top models

#ModelScoreOverall rank
1GPT-6.2 Astra42.87#4
2GPT-6.2 Sol14.59#12
3GPT-5.6 Terra14.5#72
4GPT-5.6 Sol9.69#36
5GPT-5.6 Luna1.66#133

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #5 Gemini 4 Carbon.

Interactive version: theaggregate.ai/benchmark?slug=llm-stats-nanogpt · How It Works · Data refreshed daily, snapshot 2026-11-20.