Gemini 2.5 Flash
Gemini 2.5 Flash by Google DeepMind is a hybrid reasoning model with switchable thinking, configurable thinking budgets, and native image and audio output.
Model Overview
Capabilities, design details, and architectural traits
Gemini 2.5 Flash - Google's First Fully Hybrid Reasoning Model
Gemini 2.5 Flash is explicitly described in its model card as Google's first fully hybrid reasoning model. This means thinking can be switched on or off per request, and developers can set thinking budgets to control the tradeoff between output quality, cost, and latency. Its architecture is sparse mixture-of-experts (MoE), which activates only a subset of parameters per input token, decoupling total model capacity from per-token serving cost.
Native Output Modalities That Extend Beyond Text
| Trait | Detail |
|---|---|
| Hybrid reasoning with switchable thinking | Thinking can be enabled or disabled; thinking budgets let developers tune quality-cost-latency balance per request |
| Gemini 2.5 Flash Image | A native image output variant within the same model; supports text-to-image generation, prompt-based image editing, multi-image fusion, and character/style consistency |
| Gemini 2.5 Flash Audio | A native audio output variant; supports live conversational agents, complex conversational workflows, and live speech-to-speech translation |
| 1M-token context window | Accepts text, images, audio, and video as input; text output up to 64K tokens |
Intended Use Position Within the 2.5 Family
The model card explicitly names three intended use cases for Gemini 2.5 Flash: cost-efficient thinking, well-rounded capabilities, and agentic tool use. This positions it between Flash-Lite (high-volume, low-latency tasks) and 2.5 Pro (maximum reasoning depth), with the hybrid thinking toggle as the primary mechanism for developers to adjust where on that spectrum each request lands.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Anthropic | Anthropic | |
| Release Date | May 20, 2025 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Jan 2025 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1.0M Best Context Window | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.30 Best Input Pricing | $4 | $2 |
| Output Pricing | $2.50 Best Output Pricing | $20 | $10 |
| Modalities | |||
| Inputs | fileimagetextaudiovideo | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 9.9 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | - | - | - |
| Agentic Index | - | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from Google
Other models by Google
Gemini 4 Argon
Gemini 3.8 Flash
Gemini 3.7 Flash
Gemini 3.6 Flash
Top AI Models
Leading alternatives by intelligence score
