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

A video describes Jack Wotherspoon's setup where the Interactions API calls Gemini 4 Argon to run multi-agent swarms in remote cloud sandboxes for SEC filings, Python analysis and auto-merging PRs.

The author plans an hourly loop where GPT reviews code, Gemini 4 Argon filters real issues, and Gemini 3.8 Flash fixes them, testing whether Argon can suppress GPT's false positives.

The Google Pro subscription bundled with Pixel grants access to Gemini 4 Argon, Opus 5.5 and Sonnet 5.5, though actual routing varies per account, e.g. Argon maps to Argon Trial and 3.8 Flash splits into Control and Flash-n.

Sundar Pichai says Gemini 4 Argon processed almost 25 trillion tokens in its first 72 hours, about 3x Gemini 3.0, and that Google is capacity constrained.

The author used Blender for 3D models and Antigravity with the Argon 4 model to build a flight simulator with realistic physics and challenging landings, sharing a video for ratings.

Google's Gemini business agent gains an Auto mode that routes tasks across Gemini 4 Argon, Gemini 3.8 Flash and manually selectable Claude Opus 5.5/Sonnet 5.5 to cut costs.

A same-prompt commercial platform build found Argon weaker on UI polish, domain content and code stability than Opus 5.5, but faster, cheaper and with larger context.

The same Blender MCP task, modeling an 1800s artisanal pistol, is priced across models: Gemini 4 Argon at $1.03, Claude Opus 5.5 at $1.62, GPT-6 Astra at $2.58 and Kimi K3 at $0.75.

Google's new frontier model Gemini 4 Argon is now available in Gemini for Business on Google Cloud, with Bindu Reddy saying he will move all Sol and Sonnet workloads to Argon.

A short question post asking whether Gemini 4 Carbon is expected to outperform Gemini 4 Argon, reflecting confusion over the two Gemini 4 variants.

An image post argues the Gemini 4 Argon hype is justified, pointing to benchmark numbers as evidence.

A roundup of unannounced models says Gemini 4 Flash appeared in Google's SDK and Gemini 4 Carbon is being tested on Arena, alongside GPT-6.1 Terra and Fable 5.5.