Gemini 3.8 Flash API Pricing Calculator & Benchmarks

Model Details

Released

Sep 2, 2026

Context Length

1,048,576

Max Output

65,536

Modalities

Text Audio Image Video File → Text

Capabilities

Tool use Structured outputs Prompt caching Code execution Computer use

Where is Gemini 3.8 Flash a perfect fit?

Gemini 3.8 Flash is Google’s most capable Flash model, combining advanced reasoning, coding, multimodal understanding, iterative tool use, and autonomous agent capabilities with Flash-level speed and cost efficiency.
- Long-horizon software engineering
- Autonomous coding agents
- Multi-file codebase refactoring
- Complex multi-step reasoning
- Enterprise automation
- Financial and legal analysis
- Agentic workflows with iterative tool calls
- Long-context document and data analysis
- Browser/computer-use agents
- High-volume production workloads
Gemini 3.8 Flash is specifically optimized for long-running coding and autonomous agent workflows rather than merely maximizing low-latency throughput. Google reports substantial gains over 3.7 Flash in software engineering, agentic tasks, and specialized professional reasoning.

Quick Model Estimate

Tokens Characters Words

Model

Provider

Modality

Service Tier

Context Size

Input Tokens

(USD 0.7500 per 1M tokens)

Output Tokens

(USD 3.7500 per 1M tokens)

No. of API Calls

Your Gemini 3.8 Flash Cost Estimate

💰 Total Cost

—

for 1000 input + 1000 output tokens

📥 Input (1000 × $0.750000) —

📤 Output (1000 × $3.750000) —

Cost Breakdown

📥 Input 📤 Output

Prices are watched for changes and checked against the provider's own page.

Pricing

Provider ↕	Modality ↕	Service Tier ↕	Input Price (per 1M tokens) ↕	Output Price (per 1M tokens) ↕	Cached Input (per 1M tokens) ↕	Context Size ↕	View
Perplexity	Text	Standard	$0.7500	$3.7500	$0.0750	1,048,576 tokens	→

Benchmarks

Scores from standardized evaluations by Artificial Analysis and Design Arena. Higher is better — the indices summarize overall ability, while the detailed scores break down performance on individual benchmarks.

Artificial Analysis

Higher is better · benchmarked by Artificial Analysis

Index scores by variant

Detailed scores

Variant

40.9

Intelligence Index

Overall intelligence across reasoning, knowledge & math evals

Better than 55% of 32 models

76.3

Coding Index

Coding ability across software-engineering evals

Better than 67% of 16 models

40.2

Agentic Index

Tool use & multi-step agent task performance

Better than 25% of 13 models

Reasoning

GPQA Diamond 95.3%

Graduate-level questions in biology, chemistry & physics

Humanity's Last Exam 47.8%

Expert-level questions across many academic domains

CritPt 18.3%

Unpublished physics research reasoning problems

AA-LCR 81.3%

Long-context reasoning across large inputs

Coding

SciCode 56.6%

Research-level scientific coding tasks

Knowledge

GDPval 47.3%

Economically valuable, real-world knowledge work

Omniscience Accuracy 54.6%

Breadth of factual knowledge across domains

Omniscience Non-Hallucination 44.8%

How reliably the model avoids fabricated answers

All variants

Variant	GPQA Diamond	Humanity's Last Exam	AA-LCR	SciCode	CritPt	GDPval	Omniscience Accuracy	Omniscience Non-Hallucination
Gemini 3.8 Flash (High)	95.3%	47.8%	81.3%	56.6%	18.3%	47.3%	54.6%	44.8%
Gemini 3.8 Flash (Medium)	93.5%	42.1%	84%	55.1%	12.3%	46.6%	53%	48.1%
Gemini 3.8 Flash (Low)	92%	37.1%	80.7%	55%	4%	40.8%	52.2%	35.4%

Design Arena

Elo rating by arena & category

Variant	Arena	Category	Elo	Win %	Percentile	Avg time (ms)
gemini-3.8-flash	agents	androidnative	1,084	33.4%	18	—
gemini-3.8-flash	agents	blender3d	1,232	54.3%	75	—
gemini-3.8-flash	agents	fullstack	1,244	48.1%	73	—
gemini-3.8-flash	agents	godotgamedev	1,273	53%	86	—
gemini-3.8-flash	agents	mobileapps	1,236	54.6%	85	—
gemini-3.8-flash	agents	python-pptxslides	1,138	40.3%	39	—
gemini-3.8-flash	agents	spreadsheet	1,057	41.1%	12	—
gemini-3.8-flash	agents	webapps	1,239	46.9%	67	—
gemini-3.8-flash	models	3d	1,292	49.9%	85	79,673
gemini-3.8-flash	models	asciiart	1,269	52.3%	86	57,183
gemini-3.8-flash	models	codecategories	1,302	50.4%	90	123,389
gemini-3.8-flash	models	dataviz	1,273	51.6%	82	97,674
gemini-3.8-flash	models	gamedev	1,303	51%	88	108,064
gemini-3.8-flash	models	svg	1,292	54.9%	91	56,736
gemini-3.8-flash	models	uicomponent	1,313	53.9%	90	89,142
gemini-3.8-flash	models	website	1,299	50.1%	90	136,934

Compare Gemini 3.8 Flash with

Two models side by side: worked costs, every rate, and every benchmark figure both of them report.

Deepseek v4 Pro-0813 DeepSeek Full comparison → GPT-5.6 Terra OpenAI Full comparison → GPT-5.6 Luna OpenAI Full comparison → GPT-5.4 Nano OpenAI Full comparison → GPT-5.4 Mini OpenAI Full comparison → GPT-5.4 OpenAI Full comparison →

Every comparison we publish →

FAQs about Gemini 3.8 Flash

How much does Gemini 3.8 Flash cost per 1M tokens?

Gemini 3.8 Flash costs $0.75 per million input tokens, and $3.75 per million output tokens.

What does a typical workload cost with Gemini 3.8 Flash?

1,000 requests of 2,000 input and 500 output tokens each — 2,000,000 input and 500,000 output tokens in total — costs $3.375 with Gemini 3.8 Flash at its lowest rates. Use the calculator on this page for your own volumes.

Where can I use Gemini 3.8 Flash?

Gemini 3.8 Flash is available through Perplexity, from $0.75 per million input tokens.

What is Gemini 3.8 Flash's context window?

Gemini 3.8 Flash has a context window of 1,048,576 tokens, and returns up to 65,536 tokens in a single response.

What input types does Gemini 3.8 Flash support?

Gemini 3.8 Flash accepts text, audio, image, video and file input, and returns text.