Gemini 3.7 Flash

Name: Gemini 3.7 Flash API
Brand: Google
SKU: gemini-3.7-flash

silvamux/gemini-3.7-flash

Gemini 3.7 Flash is Google's efficient model in the 3.x line, combining broad knowledge and multilingual strength with credible code reading — suited to localization, support and enterprise knowledge bases.

Supported modalities

text

Input price$1.5 / 1M tokensper 1M tokens

Output price$7.5 / 1M tokensper 1M tokens

Context1M

Max output66K

Pricing

Tune and Pro share the same prices.

Input context	Input	Output	Cache read	Cache write
-	$1.5 / 1M tokens	$7.5 / 1M tokens	$0.15 / 1M tokens	$1.5 / 1M tokens

Capabilities

Gemini 3.7 Flash scored 78 overall, performing well — an efficient model with surprising depth. It stands out in general knowledge (94%), multilingual capability (91%) and code comprehension and debugging (86%), reading and fixing code credibly while handling multilingual knowledge work, which fits localization, support and knowledge-base Q&A at volume. Comparatively, code generation (50%) is weak; writing code from scratch should go to stronger models. Fast multilingual knowledge work with light debugging is its sweet spot.

Overall

78/ 100良好

HighlightsGeneral Knowledge 94%Multilingual Capability 91% · Code Comprehension & Debugging 86%

Best forTranslation & Localization / Customer Support / Enterprise Knowledge Q&AScenario fit 82%

Relative weaknessCode Generation50%

Advanced Capabilities

8 items

Agent Collaboration73%

18 questions · Mixed

Code Generation50%

12 questions · Objective

Code Comprehension & Debugging86%

6 questions · Mixed

Creative Writing77%

22 questions · Subjective

Information Retrieval / RAG77%

9 questions · Mixed

Long-Context Handling79%

16 questions · Mixed

Math & Science Reasoning83%

14 questions · Mixed

Multilingual Capability91%

11 questions · Subjective

General Foundations

3 items

General Knowledge94%

10 questions · Mixed

Text Comprehension77%

23 questions · Mixed

Logical Reasoning67%

16 questions · Mixed

Safety & Compliance

2 items

Hallucination Control76%

15 questions · Mixed

Instruction Following79%

18 questions · Mixed

Performance

Gemini 3.7 Flash

gemini-3.7-flash

Last 24 hours · 36,230 requests

First token · TTFT · P50: 2156ms
Output speed · TPS · P50: 90tok/s
Full response · E2E · P50: 8.96s
Request success rate: 99.37%

24-hour availability

24 hours ago12 hours agoNow

Gemini 3.7 Flash example code and API

Progemini-3.7-flash

/api/v0/chat/completions/api/v1/chat/completions

UnifiedOpenAI

curl https://silvamux.io/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.7-flash",
    "messages": [{"role": "user", "content": "Hello, please introduce yourself"}]
  }'

# pip install openai
from openai import OpenAI

client = OpenAI(
    base_url="https://silvamux.io/api/v1",
    api_key="YOUR_API_KEY",
)

response = client.chat.completions.create(
    model="gemini-3.7-flash",
    messages=[
        {"role": "user", "content": "Hello, please introduce yourself"}
    ],
)

print(response.choices[0].message.content)

// Gradle: implementation(__HL_SLOT_0__)
import okhttp3.*;

OkHttpClient client = new OkHttpClient();
MediaType json = MediaType.get("application/json");
String payload = "{\"model\":\"gemini-3.7-flash\",\"messages\":[{\"role\":\"user\",\"content\":\"Hello, please introduce yourself\"}]}";

Request request = new Request.Builder()
    .url("https://silvamux.io/api/v1/chat/completions")
    .header("Authorization", "Bearer YOUR_API_KEY")
    .post(RequestBody.create(payload, json))
    .build();

try (Response response = client.newCall(request).execute()) {
    System.out.println(response.body().string());
}

// npm install openai
import OpenAI from 'openai'

const client = new OpenAI({
  baseURL: 'https://silvamux.io/api/v1',
  apiKey: 'YOUR_API_KEY',
})

const response = await client.chat.completions.create({
  model: 'gemini-3.7-flash',
  messages: [{ role: 'user', content: 'Hello, please introduce yourself' }],
})

console.log(response.choices[0].message.content)