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

A user asks whether the €5.50/month student Google AI Pro plan with Gemini 4 Argon, paired with Codex, is worth it for AI agent coding.

Opus 5.5, Sol 6.1 and Gemini 4 Argon all sit on the Pareto frontier, but Opus is more read-efficient while Argon burns more output tokens to reach the same score.

A leak claims Gemini 4 Flash runs around 300 tps, faster than GPT-6.1 Terra, with Argon-level intelligence and pricing below Argon, near Sonnet 5.5.

The author speculates Gemini 4 Flash will be extremely fast, stay cheap, focus on computer use and agentic tasks, or be scrapped in favor of another Argon version.

The post compares Argon batch at $1/$5 per 1M tokens with Opus 5.5 standard at $4/$20 and walks through a 50M-in, 10M-out migration cost example.

A leak claims Google's unannounced model codenamed Hydrogen is likely the Gemini 4 Flash tier, and if it nears Argon at a lower price it could briefly take second place from OpenAI.

On one burning-car prompt, Opus 5.5 took 20 minutes, 1.14M tokens and $3.45, while Gemini 4 Argon took 29 minutes, 2.46M tokens and $2.20, with different visuals.

Rumors around Gemini 4 Flash grow, with early Carbon testers saying its 3D output beats Argon but runs slower; the author guesses Flash pricing at $0.50–0.75 input and $2–4 output per million tokens.

The post says every lab is shipping a cheaper flagship, and Google's rumored Hydrogen could win this round if it nears Argon without burning extra tokens.

A 48-hour packet-capture test of Gemini 4 Argon finds Pro still at $15/week and Ultra at $50/5h, plus a mid-conversation cache reset that spikes single requests to $0.2–0.3.

A post rounds up rumored Gemini 4 codenames—Carbon said to target Opus 5.5 and Hydrogen close to Argon but far cheaper—and asks where that leaves OpenAI and DeepSeek V4.1 Flash.

A post claims Kimi K3.1 Pro beats Gemini 4 Argon by 5 points on coding cost-performance and notes Argon's discounted pricing ends November 30, rising to $4/$10.