gemini-4IMAGE

Gemini 4 Argon scores mid-tier on coding benchmarks, user finds

A user ran their own LLM coding benchmarks on Gemini 4 Argon and found it much better than Gemini 3.x Flash/Pro but still far behind frontier models, slow, expensive on API pricing, and unavailable on Antigravity's Pro plan.

Povilas Korop | AI Coding & Laravel Educator

@PovilasKoropArchived 21 days agoOriginally posted 2026.10.15

Gemini 4 Argon scores mid-tier on coding benchmarks, user finds

Povilas Korop | AI Coding & Laravel Educator @PovilasKoropOriginally posted 2026.10.15

Open original on X

So I ran my LLM coding benchmarks on Gemini 4 Argon. The result is... average. It's MUCH better than Gemini 3.x Flash/Pro. But still WAY behind the frontier models. Also too slow. And expensive on API pricing. And not available on Antigravity on Pro plan. Sooo over-hyped.

x.com/povilaskorop/status/2110406549755593002Archived 2026.10.15 11:36