Gemini 4 Pro / Flash: benchmarks, hands-on tests and leaks

The same prompt is run through Claude Opus 5.5, GPT-6 Astra, Grok 4.7 and Gemini 4 Argon to see which model's lobe pump rotors actually mesh and move fluid.

The author says Gemini confidently gave hours and an address for a nonexistent coffee shop without searching, so their bar for Gemini 4 Flash is simply to look things up before answering.

A roundup post previews the week's releases, including Anthropic's S-1, Big Tech earnings, and Gemini 4 Carbon and Flash.

A motion designer posts a short Duolingo-style animation made with Gemini Argon and jokes about whether designers should worry.

Google Cloud Japan's official account says Gemini 4 Argon is available via Gemini for Business, highlighting long-task context retention and expert knowledge use.

The author argues Gemini 4 Flash only needs 70% of Argon's performance plus native multimodality and free channels like Pixel, Jio and Edu to beat OpenAI 6.1 Terra and DS 4.1 Flash on price.

A reply argues the Argon rerun is only the first step toward proving Gemini 4 Argon, the Antigravity harness and Google are not falling behind.

A developer says Gemini 4 Argon performs extremely well in Antigravity, is using Gemini models across most projects again, and looks forward to the rumored Carbon model.

A user reports Argon scoring 54 with Antigravity versus 40 with the Pi harness on FrontierCS 2, calling the Antigravity setup notably better.

A user says Argon is more flexible than other models while remaining technically solid enough to handle whatever you build into an app, crediting Google.

A developer says his account was flagged by Google for potentially disrupting its services hours after he used Gemini 4 Argon agents in Antigravity to scan his own code for bugs.

Theo says Gemini hallucinates in over 80% of his review runs and calls Gemini 4 Argon one of the year's worst releases, while noting unreleased internal checkpoint Carbon.