Gemini 3.5 Flash

gemini-3.5-flash

Gemini 3.5 Flash is Google's efficient multimodal model, delivering near-Pro-level coding and reasoning at Flash-level cost and speed. It is highly optimized for coding capabilities and parallel agent execution loops, supporting text, image, video, audio, and PDF inputs. By default, it uses medium thinking effort for faster, more cost-effective responses, and fully supports thinking levels (minimal, low, medium, high) for fine-grained cost/performance trade-offs.

Context window105K

ProviderGemini

Released2026/05/20

Playground

Pricing

The same model is available through multiple service channels — choose based on latency, reliability and cost.

Prices in $ / 1M tokens

To pick a channel, add a provider field to the request body, for example "provider": { "channel": "direct" }. Valid values are direct / stable / economical; omit it to use the default channel.

Direct

Direct upstream connection — best when you need native behavior and the full context window.

Input	Output	Cache read
1.50/M	9.00/M	0.15/M

Overview

Input

Text Image

Output

Text

Gemini 3.5 Flash API: Flash-Level Cost, Pro-Level Coding Performance

Gemini 3.5 Flash is an efficient multimodal model released by Google at its I/O conference in May 2026, and the first member of the Gemini 3.5 series. Its core selling point is straightforward: delivers near-Pro-level coding and reasoning capabilities at Flash-level cost and speed.

According to Google's published data, Gemini 3.5 Flash surpasses Gemini 3.1 Pro on coding and agent benchmarks, while delivering output 4 times faster and often completing tasks at less than half the cost. It is specifically optimized for coding capabilities and parallel agent execution loops, and supports text, image, video, audio, and PDF input.

SeaWhale AI offers Gemini 3.5 Flash through an OpenAI-compatible interface, with support for tool calling, streaming output, and multimodal input.

Get an API key · Model ID: gemini-3.5-flash

Why Choose Gemini 3.5 Flash

Coding performance beyond the previous-gen Pro — outperforms Gemini 3.1 Pro on coding and agent benchmarks
4× faster output — often completes the same tasks at less than half the cost
Terminal-Bench 2.1: 76.2% — execution capability in real terminal environments
GDPval-AA: 1656 Elo — overall performance on real-world professional tasks
MCP Atlas: 83.6% — reliability in large-scale tool-calling scenarios
Full-modality input — unified processing of text, image, video, audio, and PDF

Core Capabilities

01 Coding and Parallel Agent Execution

Gemini 3.5 Flash is specifically optimized for coding capabilities and parallel agent execution loops. In agent architectures that need to run multiple subtasks simultaneously, its speed advantage is further amplified.

Multi-file code generation and modification
Parallel sub-agent execution loops
Command execution and verification in terminal environments

02 Large-Scale Tool-Calling Reliability

MCP Atlas 83.6% measures tool-calling reliability in scaled scenarios — correct parameter construction, appropriate call timing, and stable result handling. This is a key metric for production-grade agents.

Multi-tool orchestration and parallel calling
Accuracy of parameter construction
Stability in long-chain calls

03 Full-Modality Understanding

Supports text, image, video, audio, and PDF input. CharXiv Reasoning 84.2% reflects its performance on complex visual reasoning in charts and similar content.

Chart and scientific image reasoning
Video content understanding
Audio transcription and understanding
PDF document parsing

04 Adjustable Thinking Levels

Uses medium thinking effort by default for faster, more cost-effective responses, while fully supporting thinking levels from minimal to maximum, allowing compute to be precisely allocated according to task difficulty.

Default medium thinking balances speed and quality
Thinking depth can be increased for high-difficulty tasks
Simple tasks can be lowered to the minimum level

Best Use Cases

Scenario	Description
High-throughput coding assistant	Code completion and modification with fast response times
Parallel agent architectures	Execution loops running multiple subtasks simultaneously
Large-scale tool calling	Production-grade agents under tool protocols such as MCP
Chart and scientific analysis	Complex visual reasoning like CharXiv
Multimodal processing	Unified understanding of video, audio, and PDF
Cost-sensitive production deployment	Near-Pro capability at Flash pricing

Differences Between Gemini 3.5 Flash, Gemini 3.1 Pro, and 3.1 Flash Lite

Capability	Gemini 3.5 Flash	Gemini 3.1 Pro Preview	Gemini 3.1 Flash Lite
Model ID	`gemini-3.5-flash`	`gemini-3.1-pro-preview`	`gemini-3.1-flash-lite-preview`
Positioning	Efficient multimodal workhorse	Previous-gen flagship	High-capacity economy tier
Coding capability	Exceeds 3.1 Pro	Frontier	Basic
Output speed	4× vs 3.1 Pro	Standard	Fast
Terminal-Bench 2.1	76.2%	—	—
Input modalities	Text, image, video, audio, PDF	Text, image	Text, image

For specific billing, refer to the real-time pricing card at the top of the page.

How to Use the Gemini 3.5 Flash API

1. Create a SeaWhale AI API key Generate a key in the console and top up your balance.

2. Start with the default thinking level Gemini 3.5 Flash uses medium thinking effort by default, which is a suitable starting point for most tasks. Increase it only when quality is not up to standard.

3. Call the API

curl -X POST https://api.seawhaleai.com/v1/chat/completions \
  -H 'Authorization: Bearer YOUR_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "gemini-3.5-flash",
    "messages": [
      {"role": "user", "content": "Read the structure of this CSV, write a data cleaning script, and explain what each step does."}
    ],
    "stream": true
  }'

FAQ

When was Gemini 3.5 Flash released? At the Google I/O conference in May 2026; it is the first model in the Gemini 3.5 series.

Is it really stronger than 3.1 Pro? Yes, on coding and agent benchmarks, while outputting 4× faster and often at less than half the cost. In some scenarios requiring extremely deep reasoning, the Pro tier still has advantages; it is recommended to test against your own workload.

How do key benchmarks perform? Terminal-Bench 2.1 76.2%, GDPval-AA 1656 Elo, MCP Atlas 83.6%, CharXiv Reasoning 84.2%.

What input modalities are supported? Text, image, video, audio, and PDF — among the fullest modality coverage of any model listed on this page.

How do I adjust the thinking level? Default is medium. For high-difficulty reasoning tasks you can increase it; for latency-sensitive, high-frequency tasks you can lower it to the minimum.

Is it suitable for parallel agents? Very suitable. It has been specifically optimized for parallel agent execution loops, and its speed advantage is amplified in fan-out scenarios.

Why Choose SeaWhale AI for the Gemini 3.5 Flash API

Direct connection in China — no Google Cloud account or overseas network required
OpenAI-compatible interface — integrate by changing just two lines of existing code
Unified multimodal entry point — video, audio, PDF, and text share one API interface
Cross-vendor comparison — compare prices and performance side by side with the GPT and Claude series under the same account

API

API integration

Model IDUse this value as the model in inference requests

gemini-3.5-flash

API KeyBearer token used to authenticate inference requests

Manage API keys

Base URLOpenAI compatible · /chat/completions

https://api.seawhaleai.com/v2/

provider OptionalSelects a service channel; omit it and the system picks the default

"provider": { "channel": "direct" }

gemini-3.5-flash usage examples

SeaWhale AI is compatible with the OpenAI API protocol, so you can call it with the OpenAI SDK or plain HTTP requests. Streaming is enabled by default.

About the provider parameter (optional, a SeaWhale AI extension): most models are served over several channels that differ slightly in price and reliability. Add a provider field to the request body to pick one; omit it and the system selects the default channel — normal calls are unaffected.

provider is not part of the official OpenAI protocol — it is a SeaWhale AI extension that only takes effect on this platform. The OpenAI SDK allows custom fields like this to pass through; see the examples below.

Value	Channel	Best for
`direct`	Direct	The official upstream link, for native behavior and the full context window
`stable`	Preferred	Balanced availability and speed — a good default for production traffic
`economical`	Economy	Cost first, well suited to batch processing and price-sensitive workloads

Available channels and their prices are listed under "Pricing" above (channels vary by model). Additional notes:

Syntax: "provider": { "channel": "direct" }.
If the requested channel is not enabled for that model, the request automatically falls back to the default channel and returns normally — it does not error.
With the official SDKs: in Python pass it via extra_body; in Node.js put it directly on the request object and it passes through. In TypeScript projects, add a // @ts-expect-error line to skip the type check.

cURLPythonNode.js

curl https://api.seawhaleai.com/v2/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <API_KEY>" \
  -d '{
    "model": "gemini-3.5-flash",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Hello!"}
    ],
    "provider": { "channel": "direct" },
    "stream": true
  }'
# provider is optional — remove this line to use the default channel

from openai import OpenAI

client = OpenAI(
    base_url="https://api.seawhaleai.com/v2",
    api_key="<API_KEY>",
)

stream = client.chat.completions.create(
    model="gemini-3.5-flash",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"},
    ],
    stream=True,
    # Optional: pick a service channel; omit to use the default
    extra_body={"provider": {"channel": "direct"}},
)

for chunk in stream:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)

import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://api.seawhaleai.com/v2',
  apiKey: '<API_KEY>',
})

const stream = await client.chat.completions.create({
  model: 'gemini-3.5-flash',
  messages: [
    { role: 'system', content: 'You are a helpful assistant.' },
    { role: 'user', content: 'Hello!' },
  ],
  stream: true,
  // Optional: pick a service channel; omit to use the default
  // @ts-expect-error provider is a SeaWhale AI extension, not in the OpenAI SDK types
  provider: { channel: 'direct' },
})

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0]?.delta?.content ?? '')
}