Which AI model scores highest on FrontierCode v1.1 (Extended)?

GPT-6 Astra by OpenAI holds the best FrontierCode v1.1 (Extended) (extended split) result among tracked models, at 64.5% (released Sep 3 2026). Higher scores are better on this benchmark.

What are the top 3 models on FrontierCode v1.1 (Extended)?

1. GPT-6 Astra (OpenAI) — 64.5%; 2. Grok 4.6 (SpaceXAI) — 61.3%; 3. Grok 4.5 (SpaceXAI) — 56.6%.

How many models have a published FrontierCode v1.1 (Extended) score?

3 tracked models have a published FrontierCode v1.1 (Extended) (extended split) score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.

Agentic coding

FrontierCode v1.1 (Extended)extended split

frontier-difficulty agentic coding tasks (v1.1, extended split)

main split extended split

Rankings

Higher is better

FrontierCode v1.1 (Extended) — frequently asked questions

: frontier-difficulty agentic coding tasks (v1.1, extended split)
: GPT-6 Astra by OpenAI holds the best FrontierCode v1.1 (Extended) (extended split) result among tracked models, at 64.5% (released Sep 3 2026). Higher scores are better on this benchmark.
: 1. GPT-6 Astra (OpenAI) — 64.5%; 2. Grok 4.6 (SpaceXAI) — 61.3%; 3. Grok 4.5 (SpaceXAI) — 56.6%.
: 3 tracked models have a published FrontierCode v1.1 (Extended) (extended split) score. Scores are the figures reported by each lab at that model's release, so this page is a record of results over time rather than a re-run leaderboard.