Google logo

Gemini 2.5 Flash

Google · Released May 20, 2025

ProprietarybudgetcodingagentsCompare this model

$0.3

Input price / MTok

#6 of 27

$2.5

Output price / MTok

#8 of 27

1.0M

Context window

#12 of 28

222

Tokens / sec

#1 of 21

48.9

SWE-bench Verified

#12 of 12

82.8

GPQA Diamond

#13 of 19

72

AIME 2025

#10 of 10

59.3

LiveCodeBench

#7 of 8

Summary

Gemini 2.5 Flash is a proprietary language model from Google, released on May 20, 2025.

At $0.3 input and $2.5 output per million tokens, it ranks #6 of 27 for input price, placing it mid-range on input cost.

Its context window of 1.0M tokens ranks #12 of 28 for context depth in this dataset.

It targets budget, coding, agents workloads.

Throughput is measured at 222 tokens per second (ranked #1 of 21 for throughput).

First Flash model with thinking capabilities.

Benchmarks

BenchmarkScoreMax%RankSource
SWE-bench Verified48.910048.9%#12/12Google DeepMind model card (self-reporte...
GPQA Diamond82.810082.8%#13/19Google DeepMind model card (self-reporte...
AIME 20257210072.0%#10/10Google DeepMind model card (self-reporte...
LiveCodeBench59.310059.3%#7/8Google DeepMind model card (self-reporte...

Similar Models

Compare all 6 models

Models with the most shared benchmark coverage and closest scores to Gemini 2.5 Flash.

Gemini 2.5 Flash
GoogleGemini 2.5 Pro
AlibabaQwen3.7 Max
OpenAIo4-mini
xAIGrok 3
OpenAIo3
SWE-bench Verified#12 of 12
Real-world GitHub issue resolution
GPQA Diamond#13 of 19
Graduate-level science reasoning
AIME 2025#10 of 10
Advanced math competition problems
LiveCodeBench#7 of 8
Live competitive programming

Compare Gemini 2.5 Flash side by side

Pre-filled with 5 similar models

Open comparison

Speed source: https://artificialanalysis.ai/models/gemini-2-5-flash