Google logo

Gemini 2.5 Flash

GoogleProprietaryVerified

Improved across key benchmarks for reasoning, multimodality, code and long context while getting even more efficient. Best for fast performance on complex tasks.

2025-05-20
Mixture of Experts
Proprietary

Specifications

Architecture
Mixture of Experts
License
Proprietary
Context Window
1,000,000 tokens
Type
multimodal
Modalities
textimagevideoaudio

Benchmark Scores

Humanitys-Last-Exam11

A challenging benchmark of novel problems designed to test the limits of AI capabilities....

GPQA82.8

Graduate-level Problems in Quantitative Analysis (GPQA) evaluates advanced reasoning on graduate-lev...

AIME-202572

American Invitational Mathematics Examination (AIME) 2025 problems....

SWE-bench60.4

Software Engineering Benchmark (SWE-bench) evaluates models on real-world software engineering tasks...

MMMU79.7

Massive Multi-discipline Multimodal Understanding (MMMU) evaluates multimodal understanding across 3...

MMLU88.4

Massive Multitask Language Understanding (MMLU) tests knowledge across 57 subjects including mathema...

Aider-Polyglot61.9

Tests models on their ability to write code in multiple programming languages....

LiveCodeBench-v563.9

Evaluates models on their ability to solve coding problems in real-time....

SimpleQA26.9

A benchmark of simple but precise questions to test factual knowledge and reasoning....

FACTS-Grounding85.3

Evaluates models on their ability to ground responses in factual information....

Vibe-Eval65.4

Evaluates models on their ability to understand and generate content with specific vibes or styles....

Michelangelo Long-Context Reasoning (128k)74

MRCR (Multi-Round Coreference Resolution) is part of the Michelangelo benchmark suite that evaluates...

Michelangelo Long-Context Reasoning (1M)32

MRCR (Multi-Round Coreference Resolution) is part of the Michelangelo benchmark suite that evaluates...

See all benchmarks

Advanced Specifications

Model Family
Gemini
API Access
Available
Chat Interface
Available
Multilingual Support
Yes

Capabilities & Limitations

Capabilities
fast performancereasoningcodemultimodal understandinglong context
Tool Use Support
Yes

Related Models