KymaKyma
Google

Google

Gemini 3 Flash

#8 on Kyma

Google's newest Gemini on Kyma, pairing a 1M-token context window with text, image, audio, and video inputs — the model to reach for when a request needs to see, hear, or read a lot at once.

Modalities

Text+Image → Text

Below release

Not established

Weights

Not published

Where it sits in the catalogue Kyma measures

Every number against every text model Kyma prices per token — a stated rule, not a chosen line-up.

Price · input#30/70

$0.675

output $4.05 · cached $0.0675

median $0.977 · best $0.0494

Price · output#41/70

$4.05

per 1M tokens generated

median $3.531 · best $0.2164

Availability · 30d#11/68

98.62%

872 observations since 2026-05-02

median 96.2% · best 100.0%

Throughput#35/68

39.6 tok/s

probe median, one fixed prompt

median 40.8 tok/s · best 308.3 tok/s

Response time#14/68

1.42 s

probe median, to a complete answer

median 2.50 s · best 0.52 s

Context

1.05M

max output 8,192

published by Google, not measured here

Tick above each rail is this model, below it the other 69. Dashed rule is the field median, solid is its best. Better is left; the axis stops at the 90th percentile, so a few models sit past its right edge.

Usage

How much this model is actually called here.

Rank

#8

of 113 active models

Tokens served

87.1M

all-time

Platform share

2.7%

of all tokens

Tokens · last 15 daysAug 21 → Sep 4

Two clocks, and why they disagree

Kyma measures this model twice. Both are real; they answer different questions.

Probe · every 6h · 30 days

1.42s to answer

One fixed prompt, on a schedule, to every model. Comparable, because the model is the only thing that changes.

Observations
872
Answered by a substitute
3

Real traffic · last 7 days

1.24s to answer

Your requests, at the lengths clients actually send. Not comparable between models, but it is what running this one feels like.

Requests
50
Completed
92%
p95
25.3 s

Caching, as realised · last 7 days

0.0% of input cached

The share of input that actually hit cache, so the effective rate below is what was charged, not a best case.

Cached input
0
Fresh input
25.9K
List input
$0.675 /1M
Effective input
$0.675 /1M

Production prompts here are shorter than the probe's. Use the probe figure to choose between models, the traffic figure to budget for your own.

Pricing

Pay per token. Cached input bills at this model’s own cached rate, listed below.

$0.675 /1M input$4.05 /1M output
Hobby10 req/day · 2K in / 500 out
~$1.01/mo
Production1,000 req/day · 2K in / 500 out
~$101/mo
Scale20,000 req/day · 2K in / 500 out
~$2,025/mo
+ Estimate your workload
1,000
2,000
500
30%

Estimated monthly cost

$90.31

$3.01 / day on Gemini 3 Flash

Same workload on:

Gemini 3.8 Flash$137+51%
Gemini 3.5 Flash Lite$74.92-17%

Estimates use list pricing with cached input at this model's own cached rate. Actual bills depend on real token counts, and every response includes its exact cost.

When to use Gemini 3 Flash

Updated 2026-06-10

Where this model earns its cost — and where it doesn't.

Gemini 3 Flash is the newest Gemini from Google, built for long-context work and reasoning. It sits in Kyma's frontier-open quality tier and accepts text, images, audio, and video in a single request, returning text — with extended reasoning available when the problem calls for it.

On Kyma, Python apps and the OpenClaw coding agent lead its traffic. Every call gets automatic failover if a serving path degrades, and prompt caching bills repeated prompt prefixes at this model's cached input rate, which matters at this context size — resending a large cached prefix costs a fraction of the first pass.

The 1,048,576-token context window comes with function calling and structured outputs, so the long context is usable inside agent pipelines, not just for one-off summarization.

Whole-corpus analysis

The 1M-token window fits entire codebases, document sets, or transcript archives in one request instead of a retrieval pipeline.

Video and audio understanding

Send recordings, screen captures, or audio directly — no separate transcription step — and ask questions about what's in them.

Vision tasks

Screenshots, diagrams, charts, and scanned documents go in as images alongside your text prompt.

Reasoning over long inputs

Extended reasoning plus the huge context handles analysis that requires holding a lot of material in view at once.

Long-context agents

Function calling and structured outputs keep multi-step agents reliable even as the working context grows toward the 1M-token window.

Not ideal for: Single responses that need to run very long — output is capped at 8K tokens per request, so generating a book-length draft means chunking; and it sits in the premium cost tier, so high-volume simple tasks are cheaper on a lighter model.

How it compares

Against the peers people actually weigh it against.

SpecGemini 3 FlashGemini 3.8 FlashGemini 3.5 Flash Lite
Input /1M$0.675$1.013$0.405
Output /1M$4.05$5.063$3.375
Context1M1M1M
ToolsYesYesYes
ReasoningYesYesYes
Throughput39.6 tok/sNot measured57.3 tok/s

Quick start

Up and running in under two minutes.

  1. 1

    Create an API key

    Sign up and grab a key from the dashboard — $0.50 free credit, no card required.

    Get API key →
  2. 2

    Make your first request

    Drop in your key and send a chat completion — fully OpenAI-compatible.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3-flash",
    "messages": [
      {"role": "user", "content": "Explain prompt caching in one paragraph."}
    ]
  }'
  3. 3

    Stream responses

    Add "stream": true to receive tokens as they arrive.

    curl https://kymaapi.com/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3-flash",
    "stream": true,
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

FAQ

Common questions about this model.

What is the context window of Gemini 3 Flash?

Gemini 3 Flash has a 1M-token context window — roughly 1542 pages of text in a single request.

How much does the Gemini 3 Flash API cost?

$0.675 per 1M input tokens and $4.05 per 1M output tokens, with cached input at $0.0675/1M — 90% off the input rate on repeated prompt prefixes. No subscription; you pay only for what you use.

Does Gemini 3 Flash support function calling?

Yes — Gemini 3 Flash supports tool/function calling and structured outputs (JSON mode), so it works with agent frameworks out of the box.

Are the weights for Gemini 3 Flash publicly available?

No. Google does not publish Gemini 3 Flash's weights — it is reachable through an API rather than as a file you can download and run (https://deepmind.google/models/model-cards/gemini-3-flash/, read 2026-08-14). That is Google's decision about their own product, not a limitation of Kyma.

Is Gemini 3 Flash ever served below the precision its creator released it at?

Kyma cannot say, and for this model nobody can. Google does not publish Gemini 3 Flash's weights, so there is no released checkpoint whose precision a route could be measured against. We publish "not established" rather than reading a low format as a downgrade: on several models in this catalogue a low format is exactly how the creator ships it.

How do I use Gemini 3 Flash?

Kyma is OpenAI-compatible: point your SDK's base URL at https://kymaapi.com/v1, use your Kyma API key, and set the model to gemini-3-flash. Signing up is free and includes $0.50 of credit — no card required.

Can Gemini 3 Flash understand video and audio, or just images?

All three. It accepts text, image, audio, and video inputs in the same request, which makes it the model to pick on Kyma when your input isn't just text — meeting recordings, screen captures, and multimedia documents can go straight in without a separate transcription or frame-extraction step.

How does prompt caching change the cost of working with large prompts?

The first pass over a large input bills at the full input rate, but repeated prompt prefixes — the common pattern for agents and multi-turn analysis over the same corpus — bill at this model's cached input rate. Responses include the exact cost in usage.cost, so you can verify what each request actually cost.

Why use Gemini 3 Flash through Kyma?

One API key and one OpenAI-compatible endpoint covers this and every other model on the platform. You get automatic failover when a serving path degrades, plus prompt caching and $0.50 of free credit to start, no card required.

Start with $0.50 free credit — no card required.Create account →

More models by Google

See all 19 →
ModelContextInputOutput
GoogleGemini 3.8 Flash1M$1.013$5.063
GoogleGemini 3.5 Transcribe—$0.0067 / min
GoogleGemini 3.7 Flash1M$1.013$5.063
GoogleGemini 3.6 Flash1M$1.013$5.063
GoogleGemini 3.5 Flash Lite1M$0.405$3.375
GoogleGemini 3.5 Live Translate—$0.0635 / min
GoogleGemini 3.5 Flash1M$2.025$12.15
GoogleGemma 4 31B128K$0.0763$0.218