DeepSWE v1.1: Gemini 3.8 Flash scores 73.7%, beating Grok 4.7

On the DeepSWE v1.1 software engineering benchmark, Gemini 3.8 Flash scores 73.7% versus 71.0% for Grok 4.7 xHigh, with the lightweight Flash model beating xAI's flagship.

Nova

@NovaXCodeArchived 2026.09.22Originally posted 2026.09.22

DeepSWE v1.1: Gemini 3.8 Flash scores 73.7%, beating Grok 4.7 (image 1 of 2)

DeepSWE v1.1: Gemini 3.8 Flash scores 73.7%, beating Grok 4.7 (image 2 of 2)

Nova @NovaXCodeOriginally posted 2026.09.22

Open original on X

Wait… Gemini 3.8 Flash beats Grok 4.7 at Software Engineering? 🤯 DeepSWE v1.1: > Gemini 3.8 Flash: 73.7% > Grok 4.7 xHigh: 71.0% A lightweight Flash model taking down xAI’s heavy hitter in raw software engineering. This is insane!!!

x.com/novaxcode/status/2102092847604527295Archived 2026.09.22 10:19

Open original on X