Gemini 3.6 Flash
gemini-3.6-flash
Total Context
1M
Max Output
65.5K
Released
N/A
Input$1.5 / M tokens
Output$7.5 / M tokens
Gemini 2.5 Flash
gemini-2.5-flashGemini 2.5 Flash is the balanced price-performance model in the Gemini 2.5 family, combining thinking capability with lower latency and cost. Google materials position it between Pro-level depth and Flash-Lite efficiency. It is suitable for production tasks that need reasoning, multimodal input, and practical throughput.
Context Window
1M tokens
Maximum Output
65.5K tokens
Release Date
Jun 17, 2025
Modalities
Gemini 2.5 Flash Pricing
| Input Price | Output Price | Cache Read |
|---|---|---|
| $0.3/M | $2.5/M | $0.03/M |
Gemini 2.5 Flash API Capabilities
Reasoning
Supported
Tool calling
Supported
Temperature parameter
Supported
Attachments
Supported
Knowledge Base
2025-01-01
Endpoint Protocols
Completions APIMessages APIgemini
Gemini 2.5 Flash Model Highlights
Gemini 2.5 Flash combines adaptive reasoning, multimodal understanding and a large context window for responsive, high-volume workloads.
Adaptive Reasoning
The model can apply thinking when a task requires additional analysis, balancing reasoning depth with responsiveness.
Multimodal Understanding
It accepts text, images, video and audio, enabling a single model to analyze information across several media types.
Long-Context Processing
A 1,048,576-token input window supports analysis of extensive documents, media collections and consolidated application context.
Gemini 2.5 Flash Use Cases
Gemini 2.5 Flash is suited to large-scale content processing, multimodal analysis and interactive applications that still require reasoning.
Multimedia Analysis
Analyze images, recordings and videos alongside text to produce searchable summaries, classifications and extracted findings.
Content Processing at Scale
Classify, summarize and extract information from large streams of content while retaining reasoning for less routine cases.
Interactive Assistants
Build responsive assistants that interpret user content, reason through requests and return useful text responses with low perceived delay.
How to Use Gemini 2.5 Flash via the TokenHub API
import OpenAI from "openai"
const client = new OpenAI({
apiKey: process.env.TOKENHUB_API_KEY,
baseURL: "https://us-api.tokenhub.com/v1",
})
const result = await client.chat.completions.create({})
console.log(result.choices[0]?.message?.content)Gemini 2.5 Flash Benchmarks
| Index score | ||
|---|---|---|
| Artificial Analysis Intelligence Index | Artificial Analysis broad capability aggregate | 14.1 |
| Artificial Analysis Coding Index | Artificial Analysis software task aggregate | 17.8 |
| Artificial Analysis Math Index | Artificial Analysis math reasoning aggregate | 60.3 |
| Knowledge & Reasoning | ||
| MMLU-Pro | Advanced multi-task knowledge | 80.9% |
| GPQA | Advanced science problem solving | 68.3% |
| HLE | Broad expert-level exam set | 5.1% |
| Coding & Engineering | ||
| LiveCodeBench | Live coding problems | 49.5% |
| SciCode | Scientific coding challenges | 29.1% |
| Terminal-Bench Hard | Hard terminal task execution | 12.1% |
| Math | ||
| MATH-500 | Advanced math problem solving | 93.2% |
| AIME | Competition math problems | 50% |
| AIME 2025 | Competition math problems | 60.3% |
| Instruction Following & Agent Tasks | ||
| IFBench | Prompt constraint adherence | 39.0% |
| AA-LCR | Long-context reasoning | 45.9% |
| τ²-Bench | Agent workflow tasks | 14.9% |
Metrics sourced from Artificial Analysis
Frequently asked questions about Gemini 2.5 Flash
Understand what Gemini 2.5 Flash is, its best uses, distinguishing strengths, practical tradeoffs, and safe TokenHub integration guidance.
How should developers understand the role of Gemini 2.5 Flash?+
Gemini 2.5 Flash is Google’s balanced Gemini 2.5 Flash model for high-volume, low-latency tasks that still benefit from thinking. It remains a defined model generation, but newer models in the same family may be preferable for new evaluations.
When does Gemini 2.5 Flash deliver the most practical value?+
Best-fit scenarios include high-volume application requests, reliable execution of multi-step agent workflows, and analysis of text and visual inputs. Test representative inputs and define measurable acceptance criteria before production.
What are the most useful characteristics of Gemini 2.5 Flash?+
Key strengths include a strong balance of quality, speed, and cost, fast response times, and strong reasoning on difficult problems. This combination is especially useful for reliable execution of multi-step agent workflows.
What are the practical limits of Gemini 2.5 Flash?+
Consider another model when the task needs the strongest Pro-tier reasoning, the project can adopt a newer Gemini generation, or the workflow cannot include human review for important decisions. Verify important factual, legal, financial, medical, or operational outputs with qualified human review.
How should developers call Gemini 2.5 Flash through TokenHub?+
In TokenHub, select the exact model identifier displayed for Gemini 2.5 Flash, use the endpoint documented for your account, and authenticate with your TokenHub credentials. Confirm the TokenHub-exposed input types, tools, grounding options, and model lifecycle rather than assuming full Gemini API parity.
Ready to use Gemini 2.5 Flash?
Use one API key to access Gemini 2.5 Flash and more AI models through TokenHub.
Media and Discussions
Selected public videos and posts related to this model.
X (Twitter)
Reddit
YouTube