Gemini 2.5 Flash API pricing

By the aitokensai editors · Verified 2026-11-14 Sources: Gemini API pricing · Gemini API docs

Google's Gemini 2.5 Flash tier. Available at listed at platform rate (no published official USD price).

USD, standard tier. Rate read from the live rate config on 2026-11-14. Google publishes no USD list price for this model, so nothing here is cross-checked against one — this is a gateway rate, not a discount.

Specifications

What Gemini 2.5 Flash costs at your volume

Calculated at $0.15 input / $1.251 output per 1M tokens, 8.3× apart on this model. The 5:1 input-to-output assumption in this table is generous, responses longer than that cost more per response than the row suggests, so treat these as floors rather than estimates. Put your own split through the cost calculator before committing a budget.

Cheap in, expensive out: how Gemini 2.5 Flash is priced

Gemini 2.5 Flash will quote you $0.15 per million input tokens, which reads as very cheap. It is. The trap is the other side of the request: output bills at $1.251, around 8.3×. A task that sends a short instruction and gets back a long answer pays most of its cost at the output rate, and this model stops being cheap there.

Where the shape works in your favour is the reverse: large input, small output. Classification of big documents, extraction from long text, embedding-adjacent pipelines, anything that swallows a lot and answers briefly. That is what $0.15 is genuinely good for.

Where Gemini 2.5 Flash sits in the index

Gemini 2.5 Flash is 58 of 231 on input rate and 100 of 231 on output rate. A gap of 42 places, because on this model the output rate is 8.3× the input rate, steeper than the index median of 4.0× An output-heavy workload therefore pushes it 42 places down the table, which is why the cheapest model per input token is routinely not the cheapest per task. Percentile is the share of the 231 models in this index that cost more than Gemini 2.5 Flash.

On an agent mix the lowest blended rate in the same index is Gemma2 9B It at $0.01 per 1M tokens, about 97.6× less than this model at the same ratio. Across the whole index, sorting by input price alone moves a model by as much as 77 places; on Gemini 2.5 Flash specifically the shift is 42 — which is what the re-ranked index exists to show.

Gemini 2.5 Flash is 58 of 231 across every model in the index; the comparison that usually settles a choice is Google's own lineup further down.

Gemini 2.5 Flash among Google's 4 models

Gemini 2.5 Flash is the cheapest of Google's 4 models in this index at $0.15 per 1M input tokens. The range inside that group runs from Gemini 2.5 Flash at $0.15 to Gemini 4 Argon at $1. A spread of 7×, and every other model in the group costs more on input than this one.

On input rate alone, stepping up to Gemini 3.7 Flash costs 2.5× the input rate. Both measured at 2026-11-14, the same check as every other rate on this page.

Full Google range, including models outside this shortlist: Google rate index.

Similarly priced alternatives

Blended comparison (input + output per 1M). Price is one axis These are the three closest to Gemini 2.5 Flash in blended rate, not interchangeable substitutes for it.

Calling Gemini 2.5 Flash

The identifier you send is gemini-2.5-flash. Copy it exactly, because a mismatch returns a model error rather than a connection error, which sends most people to debug the wrong thing. Everything else is a base URL argument (Gemini 2).5 Flash's own SDKs work unchanged.

from openai import OpenAI

client = OpenAI(base_url="https://aicomp.ai/v1", api_key="sk-...")

resp = client.chat.completions.create(
    model="gemini-2.5-flash",
    messages=[
        {"role": "system", "content": "Answer in one short paragraph."},
        {"role": "user", "content": "Explain what changed in this release."},
    ],
    max_tokens=300,
)

# prompt_tokens / completion_tokens 是真实账单的唯一依据 —— 记下来，别猜
print(resp.usage.prompt_tokens, resp.usage.completion_tokens)

Log the usage object on every call, not just while you are optimising. On Gemini 2.5 Flash output costs 8.3× what input costs, so the same monthly token count bills very differently depending on how much of it comes back as output — measure your own split before extrapolating from the headline input rate, because rates move with vendor promotions.

Setting it up

Pointing an existing integration at Gemini 2.5 Flash is a base URL change, not a rewrite. Pick your tool:

• Claude Code — set the base URL via environment variables
• Cursor — override the API endpoint in settings
• OpenAI / Anthropic SDK — one base_url argument
• Python, sync, streaming and async examples
• Cutting the bill — eight levers ranked by effect

Where Gemini 2.5 Flash bills from: the output half

Quoting a monthly token figure tells you very little about Gemini 2.5 Flash. Output here costs 8.3× what input does, above the index median of 4.0× across 231 models, so two deployments moving the same number of tokens can differ by several times depending purely on how much the model writes back. Generation is the bill; everything else is rounding. The five shapes below are the ones Gemini 2.5 Flash actually gets used for, priced out at each one's real ratio.

The cheapest comparison is GPT-5 nano at $0.025 / $0.2 per 1M — 6.2× cheaper than Gemini 2.5 Flash on a blended rate. Treat it as a rate floor, not a recommendation: it says nothing about whether GPT-5 nano can do the work you need Gemini 2.5 Flash for.

The spread between the first and last row above is not caused by doing more work — it is caused by asking Gemini 2.5 Flash to write more. At 8.3×, the cheapest intervention available to you is shortening responses: a tighter system prompt, a lower max_tokens, a shorter answer format. Optimising the input side barely registers against your $1.251 per 1M output rate.

FAQ

Get Gemini 2.5 Flash access