Gemini 4 Neon API price compared: Google, OpenRouter, Vercel, Requesty and a marketplace (November 2026)
Disclosure: I run MartLLM (formerly UberLLM), one of the sellers in this table. Every other price was read on the seller's own pricing page or public price list on November 21, 2026. Links are at the bottom so you can ch
Disclosure: I run MartLLM (formerly UberLLM), one of the sellers in this table. Every other price
was read on the seller's own pricing page or public price list on November 21, 2026. Links are at
the bottom so you can check each one.
Gemini 4 Neon reached the API this week with a 1M-token context window and no free tier. Here is
what one million tokens costs everywhere I could find a public price, and what changes the bill.
Standard price, USD per 1M tokens
| Seller | Input | Output | Cached input |
|---|---|---|---|
| Google AI Studio / Gemini API (official) | $3.00 | $15.00 | $0.15 |
| OpenRouter | $3.00 | $15.00 | $0.15 |
| Vercel AI Gateway | $3.00 | $15.00 | $0.15 |
| Requesty (Vertex) | $3.00 | $15.00 | $0.15 |
| Requesty / Vercel, pinned to an EU or US region | $3.30 | $16.50 | $0.165 |
| MartLLM, listed price | $2.646 | $13.23 | $0.1323 |
| MartLLM, best seller offer when I looked | $1.90 | $9.50 | $0.095 |
The gateways pass Google's price through unchanged. The real differences are below.
What changes the bill
- Waiting. Google lists Batch and Flex at half price: $1.50 input, $7.50 output, $0.075 cached. OpenRouter lists the same batch price. If your job can wait, this beats every row above.
- Speed. Google's Priority tier is $5.40 / $27.00 (cached $0.27). Vercel lists the same priority price.
- Region. Pinning traffic to an EU or US region costs 10% more on Requesty and Vercel.
- Caching. Google lists implicit caching only (no explicit cache API) at 95% off input. Prompts that share a long prefix get it without extra work.
- Long prompts. Google lists one price for the whole 1M window; there is no separate tier above 200k tokens.
- Search grounding. Google: 5,000 free searches a month shared across Gemini 3 and newer, then $14 per 1,000. OpenRouter and Vercel list the same $14 per 1,000.
About the MartLLM rows
MartLLM is a marketplace: sellers post offers for the same model and each request goes to the
cheapest one that is serving. That gives two numbers, and I show both because one alone would
mislead you.
- The listed price is the most you pay: 12% below Google.
- The best seller offer moves through the day: 37% below Google when I looked, only while that seller has capacity.
What buyers were actually billed over the last 7 days, all calls added up: 21% below Google's
price (649 calls from 9 accounts). That is closer to the listed price than to the best offer, so
plan on that number, not the best one.
Adding credit by card costs 5.5% + $0.50. Count that in, and compare it with Google's Batch price
if you do not need answers right away.
Which one to pick
- You can wait minutes to hours: Google Batch or Flex at half price. Nothing here beats it.
- You need Priority speed or Google's own terms: Google direct.
- You want one key for many makers and do not mind paying list: OpenRouter, Vercel or Requesty.
- You want real-time calls below list and can live with a price that moves: a marketplace. Set a spending cap on the key and check your first invoice against the listed price.
Prices move. If you spot one that changed, comment and I will update the table.
Sources (read 2026-11-21)
- Google: https://ai.google.dev/gemini-api/docs/pricing (redirects to https://aistudio.google.com/docs/pricing), section "Gemini 4 Neon"
- OpenRouter public model list: https://openrouter.ai/api/v1/models (google/gemini-4-neon and google/gemini-4-neon:batch)
- Vercel AI Gateway public model list: https://ai-gateway.vercel.sh/v1/models (google/gemini-4-neon)
- Requesty public model list: https://router.requesty.ai/v1/models (vertex/gemini-4-neon, @eu, @us)
- MartLLM: https://martllm.com/models/google/gemini-4-neon , public list https://api.uberllm.dev/v1/models
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.