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

The author posts a video putting Gemini 4 Pro, GPT-6.1 Sol Max and Sonnet 5.5 Max side by side, apparently as a model comparison.

The poster notes that Claude Opus 5.5 and GPT-6 Sol have both shipped while Gemini 4 Pro is still unreleased, and wonders whether Google is cooking something big.

A user lists the best models right now as GPT-6.1 Sol, Claude Opus 5.5, Grok 4.7 and Gemini 4 Pro, and asks which one people would pay for if they could pick only one.

A stance post saying the author would switch to Anthropic if Fable 5.5 performs well, to Elon if Grok 5 is good value, and would simply be happy if Gemini 4 Pro launches.

A video comparison of the same prompt across three models finds Gemini 4 Pro best on depth and detail, Sonnet 5.5 solid on the candle, and Sol falling short this time.

An account shares a screenshot of what it says are Gemini 4 Pro benchmarks, claiming the AI race would get much more interesting if real, with no official source given.

A user complains that Gemini 4 Pro has drawn a flood of benchmark hype with no substance before any official release, and urges Google to just launch it.

Replying to Bindu Reddy, the user points out that Gemini 4 Pro is already on its third Arena checkpoint and questions his duopoly claim.

The author says a model labeled Gemini 3.8 Pro (28-Sep checkpoint) on Arena appears to be Gemini 4 Pro, and that it beat GPT-6.1 Sol Max in his tests.

A month-by-month list of September releases including Claude Fable 5.1, Gemini 3.8 Flash and GPT-6 Astra/Sol, with a note that a model picked on Aug 31 is already outdated.

The author generated a video on Arena with a model labeled Gemini-3.8-Flash, which he suspects was routed to Gemini 4 Pro, and calls the result scary.

A short post mocking how model verdicts swing back and forth, stringing together Gemini 4 Pro's comeback, OpenAI DevDay doom and Sol 6.1's praise as a mystical loop.