Use one EvoLink API key to connect to Gemini 4 Neon via Gemini’s native API or OpenAI-compatible Chat Completions. Check token rates and compare the cost of a completed task with Hydrogen and Carbon.
Thinking levelsImage inputFunction callingGoogle SearchGemini native API
Production routeThis rate reflects platform-side availability — only confirmed server errors (HTTP 500 / empty response) count as failures. User-side issues (content moderation, invalid params, cancellation) plus rate limits, timeouts and auth errors are excluded. Before real traffic arrives, empty buckets may display as available.Live
Copy the Gemini 4 Neon model ID, check current rates, and connect with your existing EvoLink API key.
Gemini 4 Neon
Gemini 4 model with four thinking levels
Selected
Model ID
gemini-4-neon
Best for
Start with text classification, short replies and field extraction. Before switching production traffic, compare Neon with Hydrogen or Carbon on the same tasks and record accepted results, retries and total cost.
Input
$2.850 / 1M193.8 cr / 1M
Cache read
$0.143 / 1M9.7 cr / 1M
Output
$14.250 / 1M969 cr / 1M
All rates are per 1M tokens, shown in USD and credits, and reflect your account's current pricing.
Gemini 4 Neon pricing
Estimate request costs using EvoLink token rates. Actual charges depend on usage and your account pricing.
Request calculator
Enter the token mix for one request and the number of Google Search queries.
Estimated request cost
Gemini 4 Neon
USD$0.0072
Credits0.4845
Input tokens0.1938 cr
Cache read tokens0 cr
Output tokens0.2907 cr
Google Search0 cr
Minimum charge: 0.01 credits per request.
Budget guide
Approximate requests using the current token mix and search queries.
Gemini 3.x bills each nonempty search query, even if no sources are returned. One request may run multiple queries. Search fees are added separately after the token minimum charge.
Google Search$0.014/ query·0.952 cr / query
What is the Gemini 4 Neon API?
The Gemini 4 Neon API on EvoLink accepts text, images, audio, video and documents, and returns text, structured JSON or function calls, with low, medium, high or xhigh thinking effort per request. Start with tasks your application can check, such as ticket routing, information extraction and summaries. Use the model ID gemini-4-neon with Gemini’s native API or OpenAI-compatible Chat Completions. The same EvoLink API key works across models on the gateway.
For more demanding work, evaluate Neon on a multi-file code change or ask it to extract issues from a timestamped screen recording transcript. Check whether the changes pass your tests or the reported timestamps match the recording. Use Hydrogen as a baseline and measure cost per accepted result, including retries and tool calls.
Gemini 4 Neon API specs and capabilities
The fields below describe the EvoLink route configuration. Validate model behavior on representative requests before production use.
Context window
1.05M tokens
Max output
65.5K tokens
Input
Text + images + video + files
Output
Text · JSON (structured output) · tool calls
Reasoning
Configurable reasoning effort
Tool use
Function calling with multi-step tool sequences
Prompt caching
Automatic cache reads at a lower rate
Server-side tools
Google Search, billed per search query
Protocols
Chat Completions · Gemini native API
Model ID
gemini-4-neon
Gemini 4 Neon use cases
Start with bounded, high-volume tasks where an answer can be checked automatically. Keep application-side validation and human review where an error would affect a user or a business record.
Gemini 4 Neon for classification and routing
Label tickets, messages or documents against a fixed set of categories and send each item to the right queue. Give the allowed labels and a rule for uncertain items, compare the output with a labelled sample, and track agreement and cost per accepted label.
Gemini 4 Neon for short replies and summaries
Draft short answers, notifications or summaries from text you supply. Keep the source text in the prompt, require the reply to stay within it, and measure acceptance, edits and response time before sending replies without review.
Gemini 4 Neon for routine text extraction
Pull fixed fields such as names, dates or amounts out of plain text into a format your application checks. Validate types and required fields before saving a record, and send uncertain items to review instead of an automatic retry loop.
Gemini 4 Neon for multi-step workflows
Give your coding agent a change that spans several files. Have the agent pass relevant source text to Neon, run the tests and use the results to revise the patch. For recording review, start with a timestamped transcript and check the reported issues against the footage. Track first-pass acceptance, retries and review time.
Gemini 4 Neon API integration
Use the EvoLink API from your application, or ask an Agent with terminal or HTTP access to make the request. Both use the same EvoLink API key, balance, model ID and request history.
Option 1
Call from your application
Best for: product backends, batch jobs, automated pipelines
Send text requests to the Gemini native generateContent endpoint with the model ID, system instruction and output budget.
View setup steps
1Create an EvoLink API key in the console
2Call the generateContent endpoint on the EvoLink base URL with the model ID shown above in the request path
3Send one representative request and read usageMetadata for input, cached, and output tokens
4Set an output budget and retry limit per task; validate the returned text before accepting it.
Best for: coding, review, and analysis tasks in Codex, Claude, and Gemini
Give the Agent your task, source text and acceptance criteria. It builds the API request and reports the answer and token usage.
View setup steps
1Set EVOLINK_API_KEY in your local environment; never put it in code or a prompt
2Describe the task, the inputs to include, and the expected output format
3Ask the Agent to call Gemini 4 Neon through EvoLink and show the request before sending
4Let the Agent report the answer, token usage, and any error body
Gemini 4 Neon API code example and error handling
This example shows the shortest runnable request: a Gemini native generateContent call with a system instruction, a user message, and an output budget. Open the API tab for the complete parameter and response reference.
# Gemini native API
curl -X POST "https://direct.evolink.ai/v1beta/models/gemini-4-neon:generateContent" \
-H "Authorization: Bearer $EVOLINK_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"systemInstruction": { "parts": [{ "text": "Classify the ticket as billing, bug, or other. Reply with the label only." }] },
"contents": [
{ "role": "user", "parts": [{ "text": "I was charged twice for my November invoice." }] }
],
"generationConfig": {
"maxOutputTokens": 256,
"thinkingConfig": { "thinkingLevel": "low" }
}
}'
# Set thinkingLevel to low, medium, high, or xhigh (default: high). Thinking tokens are billed at the output rate.
# For streaming, call streamGenerateContent?alt=sse with the same body.
# The response includes candidates[0].content.parts and usageMetadata
# (prompt, thinking, and output token counts).
Invalid request or unsupported parameter
Check the model ID, the contents array, and parameter ranges against the API reference; remove fields this route does not support.
Authentication or balance issue
Check the Authorization bearer token and confirm the available balance in the console.
Context length exceeded
Input, thinking and output share a 1,048,576-token window, and requests with more input are rejected. Reduce the input to the relevant material and retry.
Rate limited (429)
Back off and retry with jitter; batch or queue requests instead of sending parallel bursts.
Content rejected or unexpected output
Review the error body and validate the returned text against your task rules before retrying.
Gemini 4 Neon thinking tokens and cost
Thinking levels control a request; they do not set a fixed token allowance or guarantee a better answer. Thinking tokens are billed at the output rate ($14.250 per 1M tokens): Chat Completions counts them in completion tokens, and the native API reports them separately in usageMetadata. Keep the lowest level that passes your acceptance checks.
One 1,048,576-token window; up to 65,536 tokens per streaming request
Input, thinking, and output share one 1,048,576-token context window; requests with more input are rejected. A single streaming request returns at most 65,536 thinking and output tokens. To go further on the native API, set maxOutputTokens above that limit: a response that ends with the finish reason CONTINUATION carries a continuationToken, and resending the same request with that token continues the output until it ends with STOP or reaches maxOutputTokens. Chat Completions cannot continue; a response cut at the limit ends with finish_reason "length".
Why use EvoLink for the Gemini 4 Neon API?
Access Neon with one EvoLink API key, use the protocol that fits your application, control thinking and task budgets, and compare it with other models through the same gateway.
01
One API key for your existing client
The same EvoLink API key supports Chat Completions and the Gemini native protocol, so you can reuse an existing integration. For Chat Completions, set model to "gemini-4-neon" in the request body. For the native API, use that ID in /v1beta/models/{model}:generateContent. The two protocols have different request shapes.
Unified access
02
Set thinking effort per request
Set reasoning_effort in Chat Completions, or thinkingLevel in the native thinkingConfig, to low, medium, high or xhigh; omitting it uses high. On Chat Completions, EvoLink sends none and minimal as low. Thinking tokens are billed at the output rate, so start low on simple tasks and keep the chosen level in your usage logs.
Thinking control
03
Manage the full cost of a task
Set maxOutputTokens (max_tokens in Chat Completions) for each request; output includes thinking tokens. Small requests are subject to a minimum charge, so batch very short items where your application allows it, and cap retries for the whole task.
Cost management
04
Compare and switch models through one gateway
Compare Neon with Gemini 4 Hydrogen, Gemini 4 Carbon or your current model on the same EvoLink gateway. Trial it on a small, representative slice that includes difficult cases, and change the route only if accepted outcomes justify the total spend. Keep a tested fallback and revert when validation fails.
Model selection
Gemini 4 Neon, Hydrogen and Carbon token rates
EvoLink
Compare the EvoLink input and output rates below. Start with Hydrogen for tasks it already passes and include Carbon for harder reasoning. Compare all three on the same inputs and acceptance criteria. Add API and tool spend across all attempts, then divide by the number of accepted results. Track human review time as well.
Use your EvoLink API key on direct.evolink.ai. Send text requests with the Gemini native API or OpenAI-compatible Chat Completions. The same key works across supported models on the gateway.
What model ID should I use for the Gemini 4 Neon API?
Use gemini-4-neon as the model field in Chat Completions or in the model path of /v1beta/models/{model}:generateContent for the native API. When switching from another EvoLink model on the same protocol, keep your key and client configuration. Check the supported request fields and limits before switching traffic.
How much does the Gemini 4 Neon API cost on EvoLink?
The EvoLink rate card shows $2.850 per 1M input tokens, $14.250 per 1M output tokens and $0.143 per 1M cached input tokens. Thinking tokens use the output rate. Google Search queries are billed separately. A minimum charge applies to small requests. The cached-input rate applies to cache hits, which depend on cache availability and the request. These are EvoLink rates. Check Google’s Gemini API pricing page for direct Google pricing.
What is Gemini 4 Neon useful for?
Start with text classification, short replies, summaries or field extraction. For a harder evaluation, use a multi-file change with a test suite or a screen recording transcript with timestamps you can verify. Compare accepted results and include thinking tokens, retries and fallback requests in the cost.
How do thinking levels affect Gemini 4 Neon pricing?
Set reasoning_effort or native thinkingLevel to low, medium, high or xhigh; Neon uses high when you omit it, and xhigh thinks the longest. On Chat Completions, EvoLink sends none and minimal as low. Thinking tokens are billed at the output rate, and a higher level does not guarantee a better answer. Numeric thinking budgets are not supported.
Which inputs, outputs and tools are supported on EvoLink?
Text, image, audio, video, and PDF input, structured JSON output, function calling, and Google Search work on both protocols. Word, PowerPoint, and Excel files work on the Gemini native API as inline data or a file URL; on Chat Completions, send them as base64 file data, because Office file URLs are not accepted there. When you send function-call history back, include the thought signatures from the previous response.
What is Gemini 4 Neon's context window and output limit?
Input, thinking, and output share one 1,048,576-token window, and requests with more input are rejected. A single streaming request returns at most 65,536 thinking and output tokens. On the native API, longer outputs continue across requests: when a response ends with the finish reason CONTINUATION, resend the same request with its continuationToken until it ends with STOP. maxOutputTokens is the total budget across those requests.
How do Gemini 4 Neon, Hydrogen and Carbon compare?
Use Hydrogen as your baseline for routine tasks and include Carbon for harder reasoning. Run the same inputs and tests on each route. A higher token rate can be worthwhile if it cuts retries or manual review enough to reduce the cost of an accepted result.
Is the Gemini 4 Neon API available on EvoLink now?
Yes. Use model ID gemini-4-neon with your EvoLink API key through the Gemini native API or OpenAI-compatible Chat Completions. Access through Google AI Studio, Vertex AI or the Gemini app is managed by Google and is separate from EvoLink.