Frontier Log

Benchmark

DeepSWE v1.1

020406080Oct 202674.1%GPT-6 Astra77.9%Gemini 4 Argon020406080Oct 202674.1%GPT-6 Astra77.9%Gemini 4 Argon

Results

ModelScoreSourceHuman-verified
Gemini 4 Argon30 Sep 2026 · Google77.9%Gemini 4 Argon: our next era of frontier intelligenceSource imageNot yet
GPT-6 Astra3 Sep 2026 · OpenAI74.1%GPT-6 Astra: A new generation of intelligenceNot yet

Benchmark

deepswe.datacurve.aiHigher is better

All benchmarks