Language models · Google

Gemini API pricing for every Gemini model

Every Gemini model Google ships to its API, 15 on Revana today, through one OpenAI-compatible endpoint. Point any OpenAI SDK at https://api.revana.si/v1, name the model, and pay per token from prepaid credit, up to 52% below Google’s list price on the newest models. A request that gets no answer costs $0.00.

Gemini 4 Argon $0.96 in / $4.80 out per 1M tokens, 52% below Google’s $2.00 / $10.00
Gemini 3.8 Flash $0.36 in / $1.80 out per 1M tokens, 52% below Google’s $0.75 / $3.75
Gemini 4 Hydrogen $2.16 in / $10.80 out per 1M tokens
Revana prices as of November 2026; Google list prices read 2026-11-16 from aistudio.google.com.

Every Gemini model and what it costs

USD per million tokens on Revana as of November 2026. Click a model for its full price sheet, cache-write rates and code. Reasoning and thinking tokens bill as output.

Gemini 4

The newest Gemini models. 2 models on Revana, USD per million tokens.

Gemini 4 prices
ModelInputCached inputOutputContextGoogle list (in / out)Revana vs list
Gemini 4 Hydrogen
gemini-4-hydrogen
$2.16$0.108$10.80128K––
Gemini 4 Argon
gemini-4-argon
$0.96$0.048$4.801M$2.00 / $10.0052% below

Gemini 3.6 to 3.8 Flash

Fast models with a context window of about 1M tokens. 3 models on Revana, USD per million tokens.

Gemini 3.6 to 3.8 Flash prices
ModelInputCached inputOutputContextGoogle list (in / out)Revana vs list
Gemini 3.8 Flash
gemini-3.8-flash
$0.36$0.036$1.801M$0.75 / $3.7552% below
Gemini 3.7 Flash
gemini-3.7-flash
$1.08$0.108$5.401.05M––
Gemini 3.6 Flash
gemini-3.6-flash
$2.16$0.216$10.801.05M––

Earlier Flash and Flash Lite

Older and lighter Flash models, for prompts already tuned to them. 6 models on Revana, USD per million tokens.

Earlier Flash and Flash Lite prices
ModelInputCached inputOutputContextGoogle list (in / out)Revana vs list
Gemini 3.5 Flash Lite
gemini-3.5-flash-lite
$0.432$0.432$3.5986–––
Gemini 3.5 Flash
gemini-3.5-flash
$2.16$2.16$12.96–––
Gemini 3.1 Flash Lite Preview
gemini-3.1-flash-lite-preview
$0.36$0.36$2.16–––
Gemini 3 Flash Preview (thinking)
gemini-3-flash-preview-thinking
$0.72$0.72$4.32–––
Gemini 3 Flash Preview (no thinking)
gemini-3-flash-preview-nothinking
$0.72$0.72$4.32–––
Gemini 3 Flash Preview
gemini-3-flash-preview
$0.72$0.72$4.32–––

Pro previews

Gemini 3 and 3.1 Pro previews, with thinking versions. 4 models on Revana, USD per million tokens.

Pro previews prices
ModelInputCached inputOutputContextGoogle list (in / out)Revana vs list
Gemini 3.1 Pro Preview (thinking)
gemini-3.1-pro-preview-thinking
$2.88$2.88$17.28–––
Gemini 3.1 Pro Preview
gemini-3.1-pro-preview
$2.88$2.88$17.28–––
Gemini 3 Pro Preview (thinking)
gemini-3-pro-preview-thinking
$2.88$2.88$17.28–––
Gemini 3 Pro Preview
gemini-3-pro-preview
$2.88$2.88$17.28–––

Revana against Google’s list price

The 2 Gemini models where both of Revana’s per-token prices are below Google’s own list. The last two columns price 1,000 requests that each read 3,000 tokens and write 500, a typical chat or support call. Google list prices from https://aistudio.google.com/docs/pricing, read 2026-11-16; Gemini 3.8 Flash’s is an introductory price until 2026-12-31, Gemini 4 Argon’s is an introductory price until 2026-11-30.

Revana vs Google list
ModelGoogle list (in / out)Revana (in / out)Below by1,000 calls at Google1,000 calls on Revana
Gemini 4 Argon$2.00 / $10.00$0.96 / $4.8052%$11.00$5.28
Gemini 3.8 Flash$0.75 / $3.75$0.36 / $1.8052%$4.13$1.98

Older Gemini models are not in this table because they cost more on Revana than at Google. Same model, same answer: Revana passes the request to Google’s model unchanged.

Match the Gemini model to the job

What a workload costs on Revana at today’s prices, worked out from the token rates above.

High-volume calls

10,000 calls that each read 2,000 tokens and write 300: Gemini 3.8 Flash: $12.60 · Gemini 3.1 Flash Lite Preview: $13.68. Gemini 3.8 Flash costs 8% less for this job. Start on the lowest-priced Flash model that passes your checks; switching is a change to the model field.

Long documents and codebases

1 question about a 500K-token document: Gemini 3.8 Flash: $0.18 · Gemini 4 Argon: $0.49. Gemini 3.8 Flash costs 63% less for this job. Ask follow-up questions with the document first and the question last, so the cache takes the repeated part.

Thinking versus standard

100 answers that read 3,000 tokens and write 2,000: Gemini 3 Flash Preview (thinking): $1.08 · Gemini 3 Flash Preview (no thinking): $1.08. A thinking version costs the same per token but writes more tokens; use it where the answer needs the steps.

Hard reasoning steps

100 hard steps that read 10K tokens and write 4K: Gemini 4 Argon: $2.88 · Gemini 3.1 Pro Preview: $9.79. Gemini 4 Argon costs 71% less for this job. Send only the hard steps to the bigger model and keep the rest on Flash.

Call Gemini with the code you already have

Two settings change: the base URL and the key. Messages, tools, response_format and streaming stay in the OpenAI format, and switching to another Gemini model, or another maker’s, is a change to the model field.

curl
curl https://api.revana.si/v1/chat/completions \
  -H "Authorization: Bearer $REVANA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gemini-3.8-flash","messages":[{"role":"user","content":"Write a haiku about routing."}]}'
Python
import os
from openai import OpenAI

client = OpenAI(base_url="https://api.revana.si/v1", api_key=os.environ["REVANA_API_KEY"])
reply = client.chat.completions.create(
    model="gemini-3.8-flash",
    messages=[{"role": "user", "content": "Write a haiku about routing."}],
)
print(reply.choices[0].message.content)
print(reply.usage.cost_usd)  # what this request cost, in USD
Node.js
import OpenAI from "openai"

const client = new OpenAI({ baseURL: "https://api.revana.si/v1", apiKey: process.env.REVANA_API_KEY })
const reply = await client.chat.completions.create({
  model: "gemini-3.8-flash",
  messages: [{ role: "user", content: "Write a haiku about routing." }],
})
console.log(reply.choices[0].message.content, reply.usage.cost_usd)

The answer is OpenAI’s chat.completion object, and usage.cost_usd is the exact amount taken from your balance. For 3,000 tokens in and 500 out on Gemini 3.8 Flash:

Response
{
  "id": "chat_3f9c…",
  "object": "chat.completion",
  "model": "gemini-3.8-flash",
  "choices": [
    {
      "index": 0,
      "message": {
        "role": "assistant",
        "content": "…"
      },
      "finish_reason": "stop"
    }
  ],
  "usage": {
    "prompt_tokens": 3000,
    "completion_tokens": 500,
    "total_tokens": 3500,
    "prompt_tokens_details": {
      "cached_tokens": 0
    },
    "cost_usd": 0.00198
  }
}

Send "stream": true for server-sent events, and "stream_options": {"include_usage": true} to get the cost in the last chunk. Every field is in the chat completions docs.

Pay only for the answers you get

  • Automatic fallback. A model with more than one route starts on its lowest-cost route; a busy or failed route hands the request to the next before any answer is sent. Each model page shows its routes.
  • Held, then charged. While a request runs, Revana holds its prompt plus max_tokens (8,192 output tokens when you set none). When it ends the hold is released and only the tokens used are charged.
  • No answer costs $0.00. If every route fails you get an OpenAI-shaped error and no charge; 503 model_busy is never billed, so a short retry is safe.
  • Backup routes stay under the hold. When a model’s listed route is busy, a backup can cost more; you are charged that route’s cost, never more than the hold, and it is always in cost_usd.

Make images and video from Claude Code, Cursor and ChatGPT

  • Connect the Revana MCP server to Claude Code, Claude Desktop, claude.ai, Cursor, VS Code, Codex CLI or ChatGPT, and your agent can quote, generate and download images and video on your credit.
  • Every language model and its price is published as llms.json and through GET /v1/models, so code can pick a model by price at run time.
  • Any tool that sends HTTP works too, such as n8n’s HTTP Request node: POST https://api.revana.si/v1/chat/completions with your key as a Bearer token.

Nano Banana and Veo on the same key

Google's image and video models run on the same Revana key through POST /v1/generations, from the same credit as your Gemini tokens. A generation that fails costs nothing.

Media models on the same key
ModelKindFrom
Nano Banana 2 Liteimage$0.02 an image
Nano Banana 2image$0.03 an image
Nano Bananaimage$0.03 an image
Imagen 4 Fastimage$0.03 an image
Gemini Omni 1.2 Flashvideo$0.024 a second
Gemini Omni 1.1 Flashvideo$0.025 a second
Veo 3.1 Litevideo$0.05 a second
Veo 3.1 Fastvideo$0.08 a second

Every image, video and audio model and its price is at revana.si/models. A generation that fails costs nothing.

From signup to your first Gemini call

  1. Create a key. Sign up and create an API key in the dashboard. It works with every language, image and video model.
  2. Add prepaid credit. Packs start at $10 and 100 credits are $1. Credit never expires; no subscription or monthly minimum.
  3. Send the sample request. Copy the curl above, set your key, and read the answer and its cost_usd.

FAQ

How does Gemini API pricing work on Revana?

Each Gemini model has a price per million input tokens, cached input tokens and output tokens, taken from your prepaid Revana credit for the tokens each request used. Thinking tokens bill as output. There is no subscription or minimum.

Which Gemini model costs the least?

Gemini 3.8 Flash at $0.36 in and $1.80 out per million tokens on Revana (prices as of November 2026).

Is Gemini cheaper on Revana than on Google?

For some models, yes: Gemini 3.8 Flash costs $0.36 in and $1.80 out against $0.75 and $3.75 (52% below); Gemini 4 Argon costs $0.96 in and $4.80 out against $2.00 and $10.00 (52% below). Google's list prices were read on 2026-11-16. Others, such as Gemini 3 Flash Preview, Gemini 3 Pro Preview and Gemini 3.1 Flash Lite Preview, cost more on Revana, so compare before you move a workload.

Do thinking versions cost more?

Not per token: a thinking version has the same rates as its standard version on Revana. It usually writes more tokens, and those bill as output, so the same question costs more.

How is Revana different from Google's own Gemini API?

Revana is an OpenAI-compatible endpoint: you call Gemini with an OpenAI SDK, pay from one prepaid balance shared with image, video and other language models, and see cost_usd on every answer. Google-only features (Live API, grounding with Search, file uploads) stay on Google's API.

Do I need a Google Cloud account?

No. A Revana key and prepaid credit are enough; there is no Google Cloud project, billing account or quota request.

Does prompt caching lower the cost?

Yes. Automatic caching works as it does at Google, and cached input is billed at the cached rate in the table, a fraction of the normal input rate.

Start with one API key

One key for every language, image, video and audio model on Revana. Prepaid credit, no subscription, and a request that fails costs nothing.