Gemini 3.8 Flash

google/gemini-3.8-flash
NewFeatured
VisionAudioToolsJSONReasoning
by Google · 2026-09-02

Gemini 3.8 Flash is Google's most intelligent Flash model with significant gains from 3.7 Flash across software engineering, agentic tasks, and multi-step reasoning.

ctx1M tokens
Max output65K
Inputtext + image + video + file + audio
Outputtext
p50 TTFT4.55 s
INPUT$0.75/ 1M tokens
OUTPUT$3.75/ 1M tokens
p50 TTFT4.55 s7d
p95 TTFT10.00 s7d
TRAFFIC102.9Mtokens / 7d

Gemini 3.8 Flash is aimed at teams that need a Google model with broad multimodal input, a long context window, a large output limit, and an OpenAI-compatible API. It fits applications such as…

Who should use Gemini 3.8 Flash?

What inputs and outputs are supported?

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
import os

from openai import OpenAI

client = OpenAI(
    base_url="https://api.orcarouter.ai/v1",
    api_key=os.environ["ORCAROUTER_API_KEY"],
)

response = client.chat.completions.create(
    model="google/gemini-3.8-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$0.750
Output / 1M tokens$3.75
Cache read / 1M$0.075
CurrencyUSD

Cost calculator

Tokens / month10MM
Input share70%%
Estimated / month $16.50 · With prompt caching $14.14

Estimate based on list price

Token & cost estimator

Input tokens: 20Cost per request: $0.001890

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

Performance

p50 TTFT
4.55 s
Output speed
286 tok/s
p95 TTFT
10.00 s
Error rate
0.46%

Public benchmarks

76.3
AA Coding
Better than 95% of models compared
#7 of 137
58.7
AA Intelligence
Better than 92% of models compared
#11 of 139
GPQA Diamond
95.3
HLE-Verified
54.9
Humanity's Last Exam
47.8
Long-Context Recall
81.0
SciCode
53.6
tau_banking
44.9
terminalbench_v2_1
87.6
Source: artificialanalysis.ai, deepmind.google

Community buzz

What developers are saying this week

Hacker News13 mentions · 7dup 13 vs the previous week

How it compares

Gemini 3.8 FlashGemini 3.1 Pro PreviewGemini 3.1 Pro Preview Custom ToolsGemini 3 Flash Preview
Input $/M$0.75$2.00$4.00$0.50
Output $/M$3.75$12.00$18.00$3.00
Context1.0M1.0M1.0M1.0M
Quality9/1010/1010/109/10
Compare side-by-sideCompare side-by-sideCompare side-by-sideCompare side-by-side

FAQ

What does Gemini 3.8 Flash cost through OrcaRouter?
OrcaRouter bills Gemini 3.8 Flash at the provider rate with zero markup: $0.75 per 1M input tokens and $3.75 per 1M output tokens. No session fee or platform fee is listed in the catalog. Because output is five times the input price, long generations will cost more proportionally than long prompts.
What are the context window and maximum output size?
Gemini 3.8 Flash has a context window of 1,048,576 tokens and a maximum output of 65,536 tokens. The context window is advertised for text, image, video, file, and audio inputs. The provider does not give a separate per-media token formula in this listing.
What are Gemini 3.8 Flash's main strengths?
Its listed strengths are a 1,048,576-token context, 65,536-token maximum output, broad multimodal input, and a 54.9 score on HLE-Verified. Those features suit expert-level question work, long-document tasks, and multimedia understanding through an OpenAI-compatible API. No additional quality claims are made in the catalog.
How does Gemini 3.8 Flash compare with other models?
The catalog record does not contain benchmark scores or prices for any alternative model. Comparisons should be made by running identical test sets through OrcaRouter and measuring accuracy, cost, and speed. Use model ID google/gemini-3.8-flash alongside other IDs and compare the fields actually returned for your workload.
What data-handling terms apply?
This catalog page does not state retention, training, or data-sharing rules. If you send data to Gemini 3.8 Flash through OrcaRouter, request content is processed to complete the transaction. Business terms for data handling are found in the Google provider terms and OrcaRouter terms. Check those documents before sending regulated or confidential data.
How do I call Gemini 3.8 Flash using an OpenAI-compatible API?
Use the OrcaRouter base URL https://api.orcarouter.ai/v1 and set the model field to google/gemini-3.8-flash. Then send a standard OpenAI-style chat completion request with your OrcaRouter API key. Existing clients that support chat.completions.create should work after changing the base URL and model name.
Does OrcaRouter add a markup to the provider rate?
No. The catalog listing says the model is billed at the provider rate with zero markup. The input and output token prices shown are $0.75 per 1M input and $3.75 per 1M output. This makes the listed per-token cost the direct provider rate.
Is prompt caching available for this model?
The model facts do not mention prompt caching, cache hit pricing, or automatic cache behavior. You should not rely on cached prompts to reduce cost. The safest cost estimate is based on full input token billing at $0.75 per 1M input tokens.
What input modalities are supported?
The model accepts text, image, video, file, and audio input. Output is listed as up to 65,536 tokens and no media output modality is documented. For image, video, file, or audio prompts, test your client's content format because the API media schema is not described in the catalog.

Embed this badge

Google: Gemini 3.8 Flash$0.75/M in4546ms p50via OrcaRouter
HTML <a href="https://www.orcarouter.ai/models/google/gemini-3.8-flash" target="_blank"> <img src="https://www.orcarouter.ai/embed/google/gemini-3.8-flash.svg" alt="Google: Gemini 3.8 Flash on OrcaRouter" /> </a>
Markdown [![Google: Gemini 3.8 Flash](https://www.orcarouter.ai/embed/google/gemini-3.8-flash.svg)](https://www.orcarouter.ai/models/google/gemini-3.8-flash)

Model card as data

GET /api/public/models/google/gemini-3.8-flashOpen
Machine-readable:/llms.txt/llms-full.txt