Flagship · $3.00 in · $15.00 out · 1,048,576 tokens
Guide note · 19 Nov 2026
Gemini 4 Neon is the newest Gemini 4 model on the feed. OpenRouter describes it as a fast, mid-priced model for coding and agentic work. The practical reason to pick it is the input list: text, images, audio, video, and files, in one model.
Analyse a photo of a receipt and a short clip of a shelf
One request can hold the photo and the clip. Ask for the total, the tax, and whether the shelf matches the receipt. That is vision plus a bit of arithmetic, which is why the modality list matters more than the coding blurb.
Gemini 4 Neon is listed at $3.00 input and $15.00 output per million tokens, with a 1,048,576-token context window.
Fix a bug from a screenshot and a repo note
The listing aims Neon at repository and terminal coding. A screenshot plus the error text is the ordinary version of that job.
Gemini 4 Neon is listed at $3.00 input and $15.00 output per million tokens, with a 1,048,576-token context window.
The tasks are editorial illustrations, not measured benchmarks. The prices in them are read from today's feed.
Everyday · $1.50 in · $7.50 out · 1,048,576 tokens
Guide note · 19 Nov 2026
OpenRouter calls Gemini 4 Hydrogen the high-efficiency workhorse of Gemini 4, for high-throughput agents and coding at a lower cost than the models above it. It accepts the same broad input list as Neon.
A long contract fits a million-token window if you send the file rather than a vague question. Ask for the parties, the dates, the renewal, and any clause that is not boilerplate. This is reading, not legal advice.
Gemini 4 Hydrogen is listed at $1.50 input and $7.50 output per million tokens, with a 1,048,576-token context window.
The agent step you run thousands of times
The listing is explicit about high-throughput agent workflows. Give Hydrogen one bounded step: extract these fields, choose a label, draft a line.
Gemini 4 Hydrogen is listed at $1.50 input and $7.50 output per million tokens, with a 1,048,576-token context window.
The tasks are editorial illustrations, not measured benchmarks. The prices in them are read from today's feed.
OpenRouter describes Gemini 3.5 Flash Lite as a high-efficiency model for subagents that do one focused task inside a larger workflow. It still accepts text, images, audio, video, and files, at a lower input rate than Gemini 4.
Merchant, date, total, tax. One photo, one schema, many receipts. That is the focused task the listing describes. Check a sample of crumpled photos before you automate the rest.
Gemini 3.5 Flash Lite is listed at $0.30 input and $2.50 output per million tokens, with a 1,048,576-token context window.
A subagent that only classifies
Hand it one label set. Do not hand it the whole workflow. The listing is about subagents inside a larger system.
Gemini 3.5 Flash Lite is listed at $0.30 input and $2.50 output per million tokens, with a 1,048,576-token context window.
The tasks are editorial illustrations, not measured benchmarks. The prices in them are read from today's feed.
For most people, start with Gemini 4 Neon, then drop to the next model on this page when the task is routine. Read the cheaper-model note under each example.
Are the examples benchmarks?
No. They are concrete tasks used to explain fit. The prices next to them are live figures from the feed. The judgment is a guide note.