Hands-on tests of Google's AI: the Gemini API, the Gemini app, NotebookLM and AI Studio. Each guide shows the setup, the real output and what it cost, so you can see what works before you build on it.
9 guides
Gemini
Gemini 4 Hydrogen costs less per token than Argon, but not per task. We ran 3 office tasks and 10 hard maths problems through the API at all 4 thinking levels, against Argon, Carbon and 3.8 Flash: full marks and the fastest answers, but it thinks even at low and cost 2 to 3.7 times as much as Argon.
Gemini
We ran 120 yes/no checks through the new Gemini Decisions API and Gemini 3.5 Flash-Lite. Same accuracy, 3 times faster, but only 1.5 times cheaper with one question per call. Ten questions on one email made it 5.4 times cheaper.
Gemini
Gemini 4 Carbon costs the same per token as Argon, but not per task. We ran 3 office tasks and 10 hard maths problems on both through the API: same scores, Carbon cost 3 to 4.5 times as much because it thinks far more. At thinking level low, it matched Argon.
Gemini
How to get a Gemini API key in Google AI Studio, which models the free tier covers, why Google shows your limits only inside AI Studio, and what our paid key revealed: no quota headers and 348 thinking tokens for a one-line answer.
Gemini
We generated three dialogue clips with Gemini Omni 1.2 Flash over the API. A 15-second 720p clip cost $1.05 and took under 90 seconds, every scripted line came out word for word in English and Vietnamese, but 360p gave a different scene and one clip added a TV logo.
Gemini
Google's free student plan is AI Pro only in the US; 23 of the 24 country pages we read offer AI Plus, which comes with the Flash model. We gave 28 college problems to the Free, Plus and Pro models to see what that costs you.
Gemini
Google now publishes fixed caps only for notebooks and sources; chat and Audio Overviews use a 5-hour allowance with a weekly limit. We checked the official pages and tried it.
Gemini
Same three office tasks through both APIs: Argon passed 15/15 for $0.00679, Opus 5.5 passed 10/11 for $0.04790, and lost its point by flagging a real error in our test data.
Gemini
We gave Gemini 4 Argon and Gemini 3.8 Flash the same three office tasks via the API. Argon was faster and cheaper ($0.00687 vs $0.01135) despite a 2.7× token price, but scored 14/15 to Flash's 15/15.
