gemini-4THREAD
Gemini 4 Argon scores 55.0% on FrontierSWE
Gemini 4 Argon scores 55.0% on FrontierSWE, close to Fable 5.1 and behind only Opus 5.5 and GPT-6 Astra, with large gains over Gemini 3.8 Flash.
Proximal
@ProximalHQArchived 28 days agoOriginally posted 2026.10.01
THREAD
Gemini 4 Argon scores 55.0% on FrontierSWE The model is close to Fable 5.1 and only outperformed by Opus 5.5 and GPT-6 Astra
This post quotes another post: x.com/googledeepmind/status/2105388084154056939
Argon outperforms Gemini 3.8 Flash by a huge margin. It saturated Lean Proof Checker and Lua Native Compiler, tasks that Gemini 3.8 barely made progress on. Overall, it improved over 3.8 in 32 out of 34 tasks, only achieving lower scores in Weather Forecast and TORCS Racing Bot
Proximal @ProximalHQOriginally posted 2026.10.01
x.com/proximalhq/status/2105400332599865827Archived 2026.10.01 10:00