Gemini 2.5 Flash
Jun 2025
Gemini 2.5 FlashCompare inon
Benchmarks of Gemini 2.5 Flash50
BenchmarkByScoreRank
Overall1
Knowledge5
Math7
Science2
Coding1
Multimodal2
Writing15
Health & law3
Documents1
Other1
Legacy · saturated or no longer reported12
No benchmark matches
Rankingin thinking mode?
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperClaude Haiku 5.5unstated+28%-76%$0.10 / $0.50Remove sampling parameters+5BetterClaude Sonnet 4.5thinking+9%+606%$0.56
OpenKimi K3unstated+22%+606%$3 / $15Sampling parameters are fixed+3
CheaperMiniMax M3unstatedOpen+15%-38%$0.30 / $1.20Context: 1M → 524K+2
CheaperMiMo v2.6 ProunstatedOpen+11%-36%$0.43 / $0.87Pass back reasoning_content+3ModelScoreCostPer taskWhat changes
Claude Haiku 5.5unstated+28%-76%$0.10 / $0.50Remove sampling parameters+5MiniMax M3unstatedOpen+15%-38%$0.30 / $1.20Context: 1M → 524K+2MiMo v2.6 ProunstatedOpen+11%-36%$0.43 / $0.87Pass back reasoning_content+3
GPT-6 Lunaunstated+11%-76%$0.10 / $0.50Function calling only with none · Chat Completions+3GPT-5 Miniunstated+7%-19%$0.25 / $2Context: 1M → 400KModelScoreCostPer taskWhat changes
Claude Sonnet 4.5thinking+9%+606%$0.56Claude Opus 5unstated+58%+1,076%$5 / $25Gemini 3.1 Pro Previewunstated+46%+429%$2 / $12Cannot use both thinking controlsGemini 4 Argonunstated+46%+371%*$2 / $10Synchronous output stops at 65,536+2Claude Fable 5unstated+39%+2,253%$10 / $50ModelScoreCostPer taskWhat changes
Kimi K3unstatedOpen+22%+606%$3 / $15Sampling parameters are fixed+3MiniMax M3unstatedOpen+15%-38%$0.30 / $1.20Context: 1M → 524K+2MiMo v2.6 ProunstatedOpen+11%-36%$0.43 / $0.87Pass back reasoning_content+3
Hy4 previewOpen+7%+47%$0.83 / $2.50No image input
GLM 5.3unstatedOpen+6%+153%$1.40 / $4.40No image input+4ModelScoreCostPer taskWhat changes
Gemini 3.1 Pro Previewunstated+46%—$2 / $12Cannot use both thinking controlsGemini 4 Argonunstated+46%—*$2 / $10Synchronous output stops at 65,536+2Gemini 3 Flash Previewunstated+39%—$0.50 / $3Gemini 3.6 Flashunstated+32%—*$0.75 / $3.75Gemini 2.5 Prounstated+25%—$1.25 / $10Gemini 3.8 Flashunstated+19%—*$0.75 / $3.75Sampling parameters are ignored+1Gemini 3.5 Flash Liteunstated+8%—$0.30 / $2.50Gemini 2.5 Flash Litethinking-15%—$0.10 / $0.40All Google models →History
7 versionsBenchmarks
all categoriesChat & Knowledge
Humanity's Last Exam12.1#160
AA Omniscience Index-29.8#150
CritPt (physics)1.1#119
AA Intelligence Index 4.3.2≈13.1#166
AA-LCR (long context) 1.165.3#140
MLCR-AA8.9#44
MathArena · BRUMO 202583.3#22
MathArena · CMIMC 202554.4#22
MathArena · SMT 202575.5#23
Vals MedCode40.4#51
Vals MedScribe83#41
GPQA Diamond79#121
AIME 2025 · AA archived73.3#61
MathArena · AIME 202570.8#24
Coding
LiveCodeBench · AA archived69.5#46
Agents & Tool Use
τ²-bench · AA31.6#146
Writing & Language
IFBench (instructions)50.3#115
COMET22 averagelab83.7
COMET22 multi-domain Zh-Enlab86.5
COMET22 multi-domain En-Zhlab82.5
COMET22 multi-domain En-Delab82.3
COMET22 multi-domain De-Enlab83.3
COMET22 WMT24lab84.4
Acceptance averagelab93
Acceptance En-Zhlab95
Acceptance Zh-Enlab94
Acceptance Zh-Xlab92.5
Acceptance X-Zhlab89.4
Excellence averagelab83
Excellence En-Zhlab91
Excellence Zh-Enlab90
Excellence Zh-Xlab76.3
Excellence X-Zhlab73.8
Vision & Documents
MMMU-Pro (vision)69.1#75
FAQ
How much does Gemini 2.5 Flash cost?›
$0.30 per 1M input tokens and $2.50 per 1M output tokens; cached input $0.03 per 1M.
Is Gemini 2.5 Flash free to use?›
There is no free API; the cheapest paid price is $0.30 per 1M input tokens.
How good is Gemini 2.5 Flash, and what is it best at?›
#159 of 334 in Chat & Knowledge, #113 of 239 in Coding, #144 of 242 in Agents & Tool Use, #73 of 195 in Writing & Language, #74 of 124 in Vision & Documents.
Is Gemini 2.5 Flash better than Mistral Large 4?›
On Vals MedCode, Gemini 2.5 Flash scores 40.4 vs 40.7 — about the same.
How can I use Gemini 2.5 Flash?›
Through the API of DeepInfra, Google, Google AI Studio.
What is the context window of Gemini 2.5 Flash?›
1M tokens, with up to 66K output tokens.
When was Gemini 2.5 Flash released? Is there a newer version?›
Released 2025-06-17. It is the current version in its line.
Is Gemini 2.5 Flash open source?›
No, Gemini 2.5 Flash is a closed model available through an API. The closest open-weights option here is Kimi K3.
Can Gemini 2.5 Flash read images?›
It accepts text and images, video as input.
What is Gemini 2.5 Flash?›
Gemini 2.5 Flash is a closed language model from Google, released Jun 2025.
What are the best alternatives to Gemini 2.5 Flash?›
Claude Haiku 5.5 (cheaper), Claude Sonnet 4.5 (better), Kimi K3 (open weights).