DataLearner logo
GE

Gemini 2.5 Flash

Reasoning modelGemini FlashGemini 2.5

Gemini 2.5 Flash

Release date: 2025-04-17Updated: 2025-04-21Views: 2,287
Live demoGitHubHugging FaceCompare
Parameters
No data
Context length
1000K
Multilingual
Supported
Reasoning ability
3/5

Gemini 2.5 Flash is an AI model published by Google Deep Mind, released on 2025-04-17, for Reasoning model, and 1000K context length, with a 1134.60 score on Creative Writing.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Gemini 2.5 Flash

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
1000K tokens
Max output length
64K tokens
Model type
Reasoning model
Modality (in / out)
Text, Image, Audio, Video → Text
Release date
2025-04-17
Model file size
No data
MoE architecture
No
Total params / Active params
No data / Not applicable
Knowledge cutoff
No data
Gemini 2.5 Flash

Open source & experience

Code license
Proprietary
Weights license
Proprietary
GitHub repo
N/A
Hugging Face
N/A
Gemini 2.5 Flash

Official resources

Paper
DataLearnerAI blog
Gemini 2.5 Flash

API details

API speed
3/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Standard
TypeConditionInputOutput
Text-$0.300/ 1M tokens$2.50/ 1M tokens
Image-$0.300/ 1M tokens
Video-$0.300/ 1M tokens
Batch
TypeConditionInputOutput
Text-$0.150/ 1M tokens$1.25/ 1M tokens
Image-$0.150/ 1M tokens
Video-$0.150/ 1M tokens
Cache PricingPrompt Cache
TypeTTLWriteRead
Text-$0.030/ 1M tokens

“—” means the modality is not billed in that direction, or the vendor has not published a price for it.

Gemini 2.5 Flash

Benchmark Results

Gemini 2.5 Flash currently shows benchmark results led by AIME 2024 (16 / 62, score 88), GPQA Diamond (195 / 462, score 82.80), GeoBench ACW (9 / 20, score 76). This page also consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

9 evaluations
Benchmark / mode
Score
Rank/total
47.74
103 / 117
ARC-AGI-1
Standard Mode
32.30
125 / 147
HLE
Standard Mode
8.40
413 / 563
HLE
Standard Mode
4.70
490 / 563
HLE
unknown
12.08
367 / 563
HLE
Thinking Mode
12.10
365 / 563
HLE
Thinking Mode
11
377 / 563
CritPt
Standard Mode
1.40
135 / 200
CritPt
Thinking Mode
1.10
147 / 200

General Evaluation

2 evaluations
Benchmark / mode
Score
Rank/total
GPQA Diamond
Standard Mode
78.30
254 / 462
GPQA Diamond
Thinking Mode
82.80
195 / 462

Common Sense

2 evaluations
Benchmark / mode
Score
Rank/total
SimpleQA
Standard Mode
25.80
30 / 47
SimpleQA
Thinking Mode
26.90
29 / 47

Coding and Software Engineer

5 evaluations
Benchmark / mode
Score
Rank/total
LiveCodeBench
Standard Mode
41.10
182 / 250
LiveCodeBench
Thinking Mode
55.40
150 / 250
SWE-bench Verified
Standard Mode
50
96 / 116
SWE-bench Verified
Thinking Mode
48.90
100 / 116
WeirdML v2
16KTools
40.95
46 / 52

Math and Reasoning

5 evaluations
Benchmark / mode
Score
Rank/total
AIME 2024
Standard Mode
88
16 / 62
AIME2025
Standard Mode
61.60
136 / 215
AIME2025
Thinking Mode
72
116 / 215
IMO 2024
Standard Mode
7.80
6 / 10
4.20
40 / 80

Writing and Creative Capabilities

1 evaluations
Benchmark / mode
Score
Rank/total
Creative Writing
Standard Mode
1134.60
84 / 106

Common Sense Reasoning

1 evaluations
Benchmark / mode
Score
Rank/total
SimpleBench
Standard Mode
41.20
64 / 93

Agent Level Benchmark

7 evaluations
Benchmark / mode
Score
Rank/total
Aider-Polyglot
Standard Mode
47.10
35 / 59
55.10
27 / 59
BALROG
Standard ModeTools
33.50
9 / 12
τ²-Bench - Telecom
Standard ModeTools
14.90
258 / 264
τ²-Bench - Telecom
Thinking ModeTools
31.60
207 / 264
Terminal Bench Hard
Standard ModeTools
12.10
178 / 244
Terminal Bench Hard
Thinking ModeTools
13.60
173 / 244

Instruction Following

2 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Standard Mode
39
223 / 282
IF Bench
Thinking Mode
50.30
156 / 282

Long Context

1 evaluations
Benchmark / mode
Score
Rank/total
AA-LCR
Standard Mode
49.90
138 / 170

Claw-style Agent Evaluation

1 evaluations
Benchmark / mode
Score
Rank/total
Pinch Bench
Thinking ModeTools
70.70
32 / 38

Multimodal Understanding

3 evaluations
Benchmark / mode
Score
Rank/total
GeoBench ACW
Standard Mode
76
9 / 20
MMMU-Pro
Standard Mode
65.50
149 / 227
MMMU-Pro
Thinking Mode
69.10
136 / 227

Long Context

1 evaluations
Benchmark / mode
Score
Rank/total
Fiction.liveBench
Standard Mode
77.80
9 / 16
Gemini 2.5 Flash

Publisher

Google Deep Mind
Google Deep Mind
View publisher details
Gemini 2.5 Flash

Model Overview

Gemini 2.5 Flash is an AI model published by Google Deep Mind, released on 2025-04-17, for Reasoning model, and 1000K context length, with a 1134.60 score on Creative Writing.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code