Gemini 3.6 Flash

google/gemini-3.6-flash
NewFeatured
VisionAudioToolsJSONReasoning
by Google · 2026-07-21

Gemini 3.6 Flash is Google's latest fast, cost-efficient model in the Gemini 3 family — a natively multimodal model that reads text, images, video, audio, and files and responds in text, with a 1M-token context window and up to 64K output tokens. It is built for teams that need frontier-adjacent quality at Flash-tier speed and price, and is a strong default across coding agents, document and media understanding, retrieval-augmented generation, and high-throughput automation. Against the previous 3.5 Flash it is markedly more token-efficient and less verbose — reaching the same or better quality while emitting fewer output tokens, which directly lowers cost and latency on long agentic runs. It supports configurable reasoning effort (so callers can dial depth against speed per request), native tool calling, and structured outputs, and it holds up strongly on long-context and agentic-coding evaluations for its tier, including 91.8% on 128k-average multi-needle retrieval and competitive scores on SWE-Bench Pro, Terminal-Bench, and OSWorld. Gemini 3.6 Flash speaks both the OpenAI chat-completions format and Gemini's native generateContent API, making it a drop-in upgrade for existing Gemini and OpenAI-compatible integrations. Use it as the everyday workhorse when you want most of the intelligence of a frontier model without paying frontier prices.

ctx1.05M tokens
Max output65.5K
Inputtext + image + video + file + audio
Outputtext
p50 TTFT1.67 s
INPUT$1.50/ 1M tokens
OUTPUT$7.50/ 1M tokens
p50 TTFT1.67 s7d
p95 TTFT5.38 s7d
TRAFFIC7.6Ktokens / 7d

Google Gemini 3.6 Flash is a large multimodal model developed by Google, offered through OrcaRouter’s API with transparent pricing (no markup). It accepts text, image, video, file, and audio inputs…

What is Google Gemini 3.6 Flash and who is it for?

How does the Flash variant differ from other Gemini models?

What input modalities does Gemini 3.6 Flash support?

Code samples

Call from any SDK

OpenAI-compatible — keep the SDK you already use

  • OpenAI SDKhttps://api.orcarouter.ai/v1
  • Gemini SDKhttps://api.orcarouter.ai
from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key="$ORCAROUTER_API_KEY",
)

response = client.chat.completions.create(
    model="google/gemini-3.6-flash",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Supported parameters

  • include_reasoning
  • max_tokens
  • reasoning
  • reasoning_effort
  • response_format
  • seed
  • stop
  • structured_outputs
  • temperature
  • tool_choice
  • tools
  • top_p

Pricing

Input / 1M tokens$1.50
Output / 1M tokens$7.50
Cache read / 1M$0.150
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $33.00 · With prompt caching $28.28

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.003780

Estimate only — actual token counts depend on the provider's tokenizer.

Performance

p50 TTFT
1.67 s
Output speed
712 tok/s
p95 TTFT
5.38 s
Error rate
25.0%

Public benchmarks

69.2
AA Coding
Better than 87% of models compared
#16 of 122
50.1
AA Intelligence
Better than 82% of models compared
#22 of 124
CharXiv Reasoning (no tools)
85.2
CharXiv Reasoning (with tools)
89.4
DeepSWE v1.1
49.0
GDM-MRCR v2 8-needle (128k avg)
91.8
GDM-MRCR v2 8-needle (1M pointwise)
54.0
GPQA Diamond
92.8
Humanity's Last Exam
38.3
Long-Context Recall
69.7
MLE-Bench
63.9
OSWorld-Verified
83.0
SciCode
52.7
SWE-Bench Pro (Public)
58.7
tau_banking
24.5
Terminal-bench 2.1 (Terminus-2)
78.0
terminalbench_v2_1
77.5
Source: artificialanalysis.ai, deepmind.google

How it compares

Gemini 3.6 FlashGemini 3.1 Pro PreviewGemini 3.1 Pro Preview Custom ToolsGemini 3 Flash Preview
Input $/M$1.50$2.00$4.00$0.50
Output $/M$7.50$12.00$18.00$3.00
Context1.0M1.0M1.0M1.0M
Quality8/1010/1010/109/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What is the cost of using Gemini 3.6 Flash through OrcaRouter?
The price is $1.50 per 1 million input tokens and $7.50 per 1 million output tokens. OrcaRouter bills at the provider rate with zero markup. There are no additional fees.
What is the context window of Gemini 3.6 Flash?
The context window is 1,048,576 tokens (approximately 1 million tokens). The maximum output is 65,536 tokens. The context includes all input modalities (text, image, video, file, audio).
What are the key strengths of this model?
Gemini 3.6 Flash offers a very large context window, supports five input modalities (text, image, video, file, audio), and achieves a strong retrieval benchmark score of 91.8 on GDM-MRCR v2 8-needle (128k avg). It is also cost-efficient as a Flash variant.
How does Gemini 3.6 Flash compare to GPT-4o or Claude 3.5 Sonnet?
It has a larger context window (1M vs 128K/200K) and lower per-token pricing ($1.50/$7.50 vs ~$2.50/$10 or $3/$15). Benchmark scores are not directly comparable, but Gemini Flash is optimized for long-context retrieval. GPT-4o and Claude may offer higher reasoning accuracy in some tasks.
How does OrcaRouter handle my data when I use this model?
OrcaRouter forwards your requests to Google’s inference servers. OrcaRouter does not train on your data or store prompts beyond what is necessary for processing and billing. Google’s data privacy policies apply. You can find details in OrcaRouter’s privacy policy.
Can I call Gemini 3.6 Flash using the OpenAI Python SDK?
Yes. Set the base URL to https://api.orcarouter.ai/v1 and the model ID to "google/gemini-3.6-flash". Use the OpenAI Python library with your OrcaRouter API key. Example: client = OpenAI(api_key="your_key", base_url="https://api.orcarouter.ai/v1") then client.chat.completions.create(model="google/gemini-3.6-flash", messages=[...]).
Does the model support streaming and function calling?
Through OrcaRouter, yes. You can set stream=true and include tools in the request. The API translates the OpenAI format to Google’s API. Test with your use case to confirm full compatibility.
What is the meaning of the benchmark score 91.8 on GDM-MRCR v2 8-needle?
It measures the model’s accuracy in retrieving multiple specific pieces of information from a long context (average 128K tokens). A score of 91.8% means the model answered correctly in 91.8% of the tested retrieval scenarios. This indicates strong performance in long-context tasks.

Embed this badge

Google: Gemini 3.6 Flash$1.50/M in1666ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/google/gemini-3.6-flash" target="_blank"> <img src="https://www.orcarouter.ai/embed/google/gemini-3.6-flash.svg" alt="Google: Gemini 3.6 Flash on OrcaRouter" /> </a>
Markdown [![Google: Gemini 3.6 Flash](https://www.orcarouter.ai/embed/google/gemini-3.6-flash.svg)](https://www.orcarouter.ai/models/google/gemini-3.6-flash)

Model card as data

GET /api/public/models/google/gemini-3.6-flashOpen
Machine-readable:/llms.txt/llms-full.txt