Google API Pricing (July 2026): Gemini 3 Flash Preview $0.50/M · Gemini 2.5 Pro $1.25/M

Verified pricing data

Google API prices as of July 2026

Source: https://ai.google.dev/gemini-api/docs/pricing

As of

July 4, 2026

Update cadence

Checked daily; published only after source verification.

Source policy

Official provider pricing pages or APIs only.

Machine-readable data

/api/pricing.json

Tracked sources

https://ai.google.dev/gemini-api/docs/pricing

Estimate monthly costs with these prices

Quick pricing answers

Cheapest input

Gemini 2.5 Flash-Lite

$0.10/M input · $0.40/M output

Cheapest output

Gemini 2.5 Flash-Lite

$0.40/M output

Best cached-input price

Gemini 2.5 Flash-Lite

$0.01/M cached input

Widest context

Gemini 2.5 Pro

1,000,000 tokens context

These answers are generated from 4 verified Google models as of July 2026. Use the calculator for workload-specific totals because output tokens and cache hit rate can change the cheapest choice.

Model	Input / 1M	Cached input / 1M	Output / 1M	Context	Max output
Gemini 3 Flash Preview	$0.50	$0.05	$3.00	n/a	n/a
Gemini 2.5 Pro	$1.25	$0.125	$10.00	1,000,000	n/a
Gemini 2.5 Flash	$0.30	$0.03	$2.50	1,000,000	n/a
Gemini 2.5 Flash-Lite	$0.10	$0.01	$0.40	n/a	n/a

Official Gemini API paid-tier standard text/image/video input and output pricing. Audio input pricing differs.

Official Gemini API paid-tier standard pricing for prompts up to 200k tokens. Google lists higher prices for prompts over 200k tokens.

Official Gemini API paid-tier standard text/image/video input and output pricing. Audio input pricing differs.

Official Gemini API paid-tier standard text/image/video input and output pricing. Audio input pricing differs.

Pricing questions

What is the cheapest Google API model in July 2026?

Gemini 2.5 Flash-Lite has the lowest verified Google input-token price at $0.10/M input tokens.

Which Google model is cheapest for generated output?

Gemini 2.5 Flash-Lite has the lowest verified Google output-token price at $0.40/M output tokens.

Where does TLDL get Google pricing data?

TLDL uses official Google pricing sources and records source URLs plus verification dates in the shared pricing dataset.

Google pricing changelog

2026-07-04

Verified Gemini 3 Flash and Gemini 2.5 family standard text pricing from the official Gemini API pricing page.

https://ai.google.dev/gemini-api/docs/pricing

TLDL tracks Google Gemini API pricing in the shared pricing dataset used by the public pricing API, the LLM cost calculator, and the main LLM API pricing comparison.

The verified table above covers the standard paid-tier token rates in TLDL's dataset. It is generated from the same JSON file as the calculator and public API, so the model prices, source URL, and verification date stay aligned across all three.

Gemini API pricing at a glance

Gemini API cost depends on four choices: the model, input tokens, output tokens, and whether the request uses standard, cached, or batch processing. Text, image, video, and audio inputs can also have different rates. The table above focuses on comparable text-token pricing; check Google's linked pricing page before estimating multimodal, grounding, or live API workloads.

Free tier vs paid tier

Google offers a free Gemini API tier for development and smaller projects, with limited model access and lower rate limits. The paid tier adds higher production limits, context caching, Batch API access, and additional models. Google states that content submitted on the free tier may be used to improve its products, while paid-tier content is not used for that purpose.

Free-tier availability and rate limits vary by model and can change independently of token prices. Treat the free tier as a way to prototype, not as a guaranteed production budget.

Standard, batch, and cached input

• Standard requests are interactive calls billed at the model's regular input and output rates.
• Batch API requests run asynchronously and Google prices them at 50% of the equivalent standard token cost. They suit evaluations, enrichment, classification, and other jobs that do not need an immediate response.
• Cached input can reduce the price of repeatedly sending the same large context. Explicit caching may also carry a storage charge based on cached tokens and time.

Output rates include thinking tokens where Google specifies that in its pricing table. That means a short visible answer can still use more billable output tokens when reasoning is enabled.

How to calculate Gemini API cost

For a standard text request, use:

cost = (input tokens / 1,000,000 × input rate) + (output tokens / 1,000,000 × output rate)

For example, if a model costs $0.50 per million input tokens and $3 per million output tokens, a workload using 2 million input tokens and 500,000 output tokens costs:

(2 × $0.50) + (0.5 × $3) = $2.50

This is a token-only estimate. Add any applicable caching storage, Google Search grounding, live API, image generation, or other feature charges. For repeated workloads, enter monthly token totals in the LLM cost calculator.

Which Gemini model is cheapest?

Use the Cheapest input and Cheapest output cards above for the current answer from TLDL's verified models. The lowest token rate is not always the lowest total cost: a stronger model may need fewer retries, fewer output tokens, or a shorter prompt to finish the same task.

As a practical starting point:

• Choose a Flash or Flash-Lite class model for high-volume extraction, classification, and latency-sensitive features.
• Choose a Pro class model when complex reasoning or coding quality matters more than the lowest per-token rate.
• Use Batch API when a turnaround of up to 24 hours is acceptable.
• Test with your own prompts and measure both input and output tokens before committing to a model.

How to use this page

Use the quick answers for the current cheapest Gemini input, output, cached-input, and context-window choices. Use the table when you need per-model input price, cached input price, output price, context window, max output, source URL, and verification date in one place.

For production billing decisions, also check the linked official Google Gemini pricing page. TLDL is designed to catch drift and make source status visible, not to replace a provider's billing terms.

Related comparisons

Gemini is often compared with GPT, Claude, DeepSeek, and budget hosted open models. Start with the provider table here, then use the full comparison or calculator when routing traffic across vendors.

FAQ

Is the Gemini API free?

Gemini has a free API tier for eligible models and regions, subject to model-specific rate limits. Production applications generally need the paid tier for higher limits and paid-only features.

How much does the Gemini API cost per token?

Google publishes prices per one million tokens rather than per individual token. Divide the per-million rate by 1,000,000 to get the unit price, then multiply by the input or output tokens used. Rates vary by model, modality, prompt length, and processing mode.

Is Gemini Batch API cheaper?

Yes. Google documents Batch API token processing at 50% of standard interactive pricing for supported models. Batch jobs are asynchronous and are designed to complete within 24 hours.

Does Gemini charge for thinking tokens?

For models whose official price lists output as including thinking tokens, those reasoning tokens are included in billable output usage. Review actual usage rather than estimating from visible response length alone.

Is Gemini API pricing the same as Vertex AI pricing?

Not necessarily. This guide tracks the Gemini Developer API. Google notes that Vertex AI pricing may differ, so Vertex deployments should be estimated from the Vertex AI pricing documentation.

Related

• LLM API Pricing Comparison
• LLM Cost Calculator
• OpenAI API Pricing
• Claude API Pricing
• DeepSeek API Pricing