Bonus: Top up now and we'll double your first deposit — get x2 credits instantly.
Get Started
Sign Up
Sign Up

Gemini 3.5 Flash API: Endpoint & Pricing

Gemini 3.5 Flash is Google’s natively multimodal reasoning model optimized for very low latency and cost while maintaining frontier‑level performance, particularly for coding and agentic workflows.

Start Using API

What is Gemini 3.5 Flash?

Gemini 3.5 Flash is a proprietary large multimodal reasoning model from Google designed to deliver fast, cost‑efficient, frontier‑level intelligence for real‑time applications. It is mainly used to power AI agents that perform complex, long‑horizon tool-using workflows, and to provide high-throughput, low-latency text, code, and multimodal generation in products like the Gemini app, AI Mode in Search, and the Gemini API. It also serves as a workhorse model for enterprise integrations where speed and scale are critical, such as agent platforms and developer tooling. Gemini 3.5 Flash belongs to Google’s Gemini model family and is an evolution of earlier Flash variants like Gemini 3 Flash and Gemini 2.5 Flash.

Can you self-host Gemini 3.5 Flash?

No. We found no public weights for Gemini 3.5 Flash on Hugging Face, so there is nothing to download, quantize or serve on your own GPUs: no VRAM budget, no vLLM or GGUF build to plan for. The only way to run it is over an API.

Skip the deploy — use LLM.API as your endpoint

No GPUs, no quantization trade-offs. Same model, OpenAI-compatible, with custom volume discounts on request.

Get API keySee API code ↓

Providers

Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick — Balanced (price + speed), Nitro (fastest), or Exacto (highest tool-calling accuracy).

List price by provider

InputOutput
google-ai-studio$1.5 in
$9 out
google-vertex$1.5 in
$9 out

Provider list prices; the LLM.API discount applies on top.

Provider Input Output Cache read /M Latency Throughput Uptime
google-ai-studioin $1.5; out $9 per 1M tokens1M tokensvision, tools, streaming, reasoning, web search, JSON
google-vertexin $1.5; out $9 per 1M tokens1M tokensvision, tools, streaming, reasoning, web search, JSON

Prices and availability from the LLMAPI catalogue, updated nightly. Last updated 11 Oct 2026.

Try this model

Test Gemini 3.5 Flash right here — free to start.

Gemini 3.5 Flash
Hi! Want to test the model?

Suggestions for your first prompt

Code snippet

Call the model through the OpenAI-compatible API.

Switching from OpenAI? Change 2 lines.

- base_url="https://api.openai.com/v1"
- api_key="YOUR_OPENAI_KEY"
+ base_url="https://api.llmapi.ai/v1"
+ api_key="YOUR_LLMAPI_KEY"

Everything else stays the same — SDK, message format, tools, streaming. Set model to gemini-3.5-flash and the call runs unchanged.

bash
curl https://api.llmapi.ai/v1/chat/completions \
  -H "Authorization: Bearer $LLMAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.5-flash",
    "messages": [
      {"role": "user", "content": "Give me three crisp launch checklist items."}
    ],
  "reasoning_effort": "minimal"
  }'
from openai import OpenAI

client = OpenAI(
    api_key="YOUR_LLMAPI_KEY",
    base_url="https://api.llmapi.ai/v1",
)

resp = client.chat.completions.create(
    model="gemini-3.5-flash",
    messages=[
        {"role": "system", "content": "You are a precise product assistant."},
        {"role": "user", "content": "Give me three crisp launch checklist items."},
    ],
    reasoning_effort="minimal",
)
print(resp.choices[0].message.content)
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.LLMAPI_KEY,
  baseURL: "https://api.llmapi.ai/v1",
});

const resp = await client.chat.completions.create({
  model: "gemini-3.5-flash",
  messages: [
    { role: "user", content: "Give me three crisp launch checklist items." },
  ],
});

console.log(resp.choices[0].message.content);
stream = client.chat.completions.create(
    model="gemini-3.5-flash",
    messages=[{"role": "user", "content": "Write a haiku about latency."}],
    stream=True,
)

for chunk in stream:
    delta = chunk.choices[0].delta.content
    if delta:
        print(delta, end="", flush=True)
{
  "model": "gemini-3.5-flash",
  "messages": [
    {"role": "system", "content": "You are a precise product assistant."},
    {"role": "user", "content": "Give me three crisp launch checklist items."}
  ],
  "reasoning_effort": "minimal"
}

ENDPOINTS

Supported endpoints & features

Base URL https://api.llmapi.ai/v1 · model ID gemini-3.5-flash · OpenAI-compatible.

  • POST /v1/chat/completionsChat Completions — supported
FeatureStatusHow to set it
Function calling (tools)Supportedtools, tool_choice
Parallel tool callsNot supportedparallel_tool_calls
JSON modeSupportedresponse_format: {"type": "json_object"}
Structured outputsNot supportedresponse_format: {"type": "json_schema", …}
StreamingSupportedstream: true
Vision (image input)Supportedimage_url content parts
Web searchSupportedweb_search
Reasoning effortminimal · low · medium · highreasoning_effort: "minimal"

ERRORS

Errors & fallback routing

CodeWhat it meansWhat to do
401Missing or invalid API key.Send Authorization: Bearer YOUR_LLMAPI_KEY and check the key is active.
429Rate limited, or the upstream provider is throttling the request.Back off and retry with jitter; the gateway also retries the request on another provider where one is available.
5xxUpstream provider error or timeout.Retry; fallback routing sends the retry to the next healthy provider for this model.

When a provider fails or throttles, the request is routed to the next provider serving this model — the ones listed in the providers table above.

Uptime

30-Day Uptime
100.00%
Past Incidents (30d)
0
Error rate (24h)
0.00%

Last 30 days

30/30 days operational | 100.00% uptime

30 days ago Today
Operational Degraded Outage Maintenance
See All Incidents

Gemini 3.5 Flash in production

What the model is good at.

  • Conversational Chat

    Handles multi-turn dialogue, follows complex instructions, and maintains context to assist with questions, brainstorming, and task guidance.

  • Code Assistance

    Generates, reviews, and explains code snippets in multiple languages, helping with debugging, refactoring, and implementation details.

  • Image Understanding

    Interprets images by recognizing objects, text, and visual relationships, enabling descriptive captions and reasoning about visual content.

  • Multilingual Translation

    Translates between many languages and can rewrite text to improve clarity, tone, and style for different audiences.

  • Optical Character Recognition

    Extracts readable text from images or screenshots containing documents, signage, or handwriting for downstream search and analysis.

What you get when you call Gemini 3.5 Flash through LLM.API.

  • Unified AI Routing

    Dynamically route each request to the best model across providers using policies, performance signals, and constraints—without changing your integration or redeploying code.

    One endpoint, every model

  • Cost-Aware Orchestration

    Optimize spend automatically by mixing premium and budget models, enforcing per-request budgets, and monitoring real-time token usage across vendors from a single control plane.

    Max performance, minimal spend

  • Resilient Fallback Logic

    Define provider and model fallbacks that trigger on errors, timeouts, or quality checks so your production workloads stay up even when vendors don’t.

    Stay online, fail safely

  • End-to-End Observability

    Trace every request across models and providers with logs, metrics, and structured payloads to debug latency, failures, and quality issues in one console.

    See every token hop

  • Task-Level Abstractions

    Ship faster with high-level tasks—chat, tools, rerank, embed—so you can swap underlying models or providers without refactoring your application logic.

    Code to tasks, not models

  • High-Throughput Batch Jobs

    Run large-scale inference jobs—embeddings, scoring, generation—through a unified batch API with concurrency controls, retries, and progress tracking built in.

    Scale jobs, not scripts

When to Use — When NOT to Use

Use it if...

  • You need a low-cost general-purpose model for high-volume API traffic and experimentation.
  • You need fast responses for chatbots, simple agents, and customer-support automations.
  • Your use case involves lightweight code generation, refactoring, and short code explanations.
  • Your use case involves summarizing short to medium documents, emails, or web pages.
  • You need multilingual understanding and translation for everyday content across many major languages.
  • Your use case involves basic image understanding, captioning, and simple visual question answering.
  • You need inexpensive fine-tuning or prompt-tuning to specialize a model for specific tasks.

Avoid if...

  • You need frontier-level reasoning and accuracy comparable to state-of-the-art flagship models.
  • Your workload requires highly reliable, domain-expert answers for medical, legal, or financial decisions.
  • You need complex multi-step tool use, planning, or large autonomous agents with long horizons.
  • Your workload requires consistently strong performance on very long-context documents or codebases.
  • You need best-in-class code generation, debugging, and architecture design for large software systems.
  • Your workload requires advanced vision reasoning over technical diagrams, dense charts, or medical images.

LLM API rating for Gemini 3.5 Flash

★ 4.2/ 5.0LLM API rating
50% production adoption · 30% category benchmark · 20% tooling & value✓ Verified benchmark · LMArena + Artificial Analysis
4.8Production adoptionVery high production volume
3.8Benchmark qualityArtificial Analysis Intelligence Index 33.6 · LMArena text rank #15, Elo 1481
3.3Developer ecosystemContext window, tool and JSON support, price against the category median
Reasoning & chat 4.2Coding 4.2Vision & multimodal 4.2Tool & function calling 4.6Long context 4.6Voice & audio 4.0
How is this rating calculated?

Every model gets a separate 1.0–5.0 score for each category it can be used in, so a model can rate high for coding and lower for vision. The score combines three parts:

  1. Production adoption — 50%. How widely the model is actually used in production: provider traffic, LiteLLM and gateway usage, open-weight downloads and its LMArena rank. Proven, widely used models score highest.
  2. Category benchmark — 30%. The public benchmark that matters for that category — the Artificial Analysis Intelligence Index and LMArena Elo for reasoning, LMArena coding Elo for coding, image and document support plus vision Elo for vision, tool and JSON support for function calling, measured speed for voice.
  3. Tooling & value — 20%. Context window, support for tools, structured outputs and streaming, and price compared with the category median.

Scores of 4.0 and above are shown in green, 3.0–3.9 in yellow, below 3.0 in red. When a model has no public data for a part, it gets a neutral value instead of an invented number.

A “Verified benchmark” mark (✓) means an independent lab — LMArena or Artificial Analysis — has actually tested that model, and the source figure is shown. Models without that mark are labelled “Estimated from published specs”: the score comes from the published context window, tool and schema support, price and adoption only, because no independent lab has tested them yet.

One exception: a brand-new top-tier model that scores very high on the independent benchmarks but has no usage history yet is rated on the benchmark (60%) and tooling & value (40%) alone, so a missing adoption figure can't hold it back. All inputs come from public sources and are refreshed weekly.

Adoption and benchmark inputs come from public sources — LMArena, Artificial Analysis Intelligence Index and published model specs. Updated weekly.

COMMUNITY

What developers say about Gemini 3.5 Flash

LMArena#15 · Elo 1481Arena coding#30 · Elo 1492Speed—Sources: LMArena, Artificial Analysis

Summarised from public developer discussions and write-ups. Opinions are the sources’, not LLM.API’s.

  • Developers on the Google AI Forum report a regression in coding quality, noting the model makes more mistakes and alters unrelated project code compared to Gemini 3 Flash.
  • Developers indicate that while Gemini 3.5 Flash is exceptionally fast and agent-oriented, it struggles with instruction following, such as failing to verify data against multiple requested sources.
  • Opinions are divided; some developers claim it handles complex systems programming and bug fixing effectively, while others argue it hits rate limits quickly and is prone to errors.

SOURCES

Frequently Asked Questions

  • What is Gemini 3.5 Flash?

    Gemini 3.5 Flash is a lightweight, multimodal Google model optimized for fast, low-cost inference on text and image tasks.

  • What modalities does Gemini 3.5 Flash support via LLM.API?

    Through LLM.API, Gemini 3.5 Flash supports text input and output, and image input with text output for vision-language tasks.

  • What is the context window of Gemini 3.5 Flash?

    Gemini 3.5 Flash supports up to a 1 million token context window, suitable for very long conversations or documents.

  • What is Gemini 3.5 Flash best suited for?

    Gemini 3.5 Flash is best for high-throughput, latency-sensitive workloads like chatbots, routing, classification, and lightweight reasoning over text and images.

  • How is Gemini 3.5 Flash priced on LLM.API?

    LLM.API exposes Gemini 3.5 Flash with per-token pricing, typically significantly cheaper than flagship reasoning models; check the LLM.API pricing page for current rates.

  • How fast is Gemini 3.5 Flash in practice?

    Gemini 3.5 Flash is designed for low latency and high throughput, generally returning responses faster than larger Gemini reasoning models at similar token counts.

  • How do I call Gemini 3.5 Flash through LLM.API?

    Set the model field to "google/gemini-3.5-flash" in your LLM.API completion or chat endpoint request, and pass prompts like with other text models.

  • How does Gemini 3.5 Flash compare to larger Gemini 3.5 models?

    Gemini 3.5 Flash is cheaper and faster but generally less capable at complex reasoning, coding, and nuanced instruction following than the larger Gemini 3.5 models.

  • Does Gemini 3.5 Flash support structured outputs via LLM.API?

    Yes, you can use JSON-style or tool-calling schemas through LLM.API, but outputs are not guaranteed to be perfectly well-formed in all cases.

  • What are the main limitations of Gemini 3.5 Flash?

    Gemini 3.5 Flash may struggle with deep multi-step reasoning, precise long-code generation, and domain-expert tasks compared to heavier, more capable models.

  • Can I fine-tune Gemini 3.5 Flash through LLM.API?

    Direct fine-tuning of Gemini 3.5 Flash is not exposed via LLM.API; instead, use prompt engineering, retrieval, and system prompts to adapt behavior.

  • Does Gemini 3.5 Flash support streaming responses on LLM.API?

    Yes, LLM.API can stream Gemini 3.5 Flash tokens incrementally, which is recommended for chat applications needing minimal perceived latency.

COMPARE

Competitive Models

Gemini 3.1 Flash TTS Preview is Google’s low-latency text‑to‑speech model that generates natural, expressive speech with fine-grained control via style prompts and audio tags. It is optimized for fast, high‑quality voice synthesis across many languages and voices.

  • Nano Banana 2 (Gemini 3.1 Flash Image Preview)

    • Text Generation

    Nano Banana 2 (Gemini 3.1 Flash Image Preview) is Google DeepMind’s image generation and editing model built on the Gemini 3.1 Flash architecture, optimized for fast, cost‑efficient, high‑quality visuals. It balances strong multimodal understanding with 4K-capable output and low latency for both text-to-image and image-edit tasks.

  • Gemini 3 Flash Preview

    • Instruction Following

    Gemini 3 Flash Preview is a Google multimodal large language model optimized for high speed and cost‑effective performance in complex reasoning tasks. It offers long‑context understanding and strong support for agents, coding, and retrieval‑augmented applications.

    • Get one key to every model

    Swap your API key. Keep your code.

    Start for free Talk to sales