Gemini 2.5 Flash Preview (Sep '25)
Sep 2025
Gemini 2.5 Flash Preview (Sep '25)Compare inon
Gemini 2.5 Flash Preview (Sep '25) measured on 23 of 104 Chat & Knowledge benchmarks23
BenchmarkByScoreRank
Overall1
Knowledge17
Science2
Documents1
Legacy · saturated or no longer reported2
No benchmark matches
—per task?
not available via APIprice
—output speed?
1Mcontext
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-5.6 Lunaxhighwins 55%—$0.085
OpenGLM 5.3maxwins 61%—$2.01No image input+4BetterGPT-5.6 Terraxhighwins 56%—$0.63Prompts with >272K input tokensBetterGPT-5.6 Solxhighwins 58%—$1.18BetterGemini 3.8 Flashhighwins 61%—*$1.24Sampling parameters are ignored+1ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-5.6 Lunaxhighwins 55%—$0.085GPT-5.6 Terraxhighwins 56%—$0.63Prompts with >272K input tokensGPT-5.6 Solxhighwins 58%—$1.18Gemini 3.8 Flashhighwins 61%—*$1.24Sampling parameters are ignored+1GLM 5.2maxOpenwins 58%—$1.47No image inputModelScoreCostPer taskWhat changes
GLM 5.3maxOpenwins 61%—$2.01No image input+4
Kimi K3maxOpenwins 60%—$2.00Sampling parameters are fixed+3
NVIDIA Nemotron 3 Ultra 550B A55B NVFP4Openwins 58%——Context: 1M → 262K+1GLM 5.2maxOpenwins 58%—$1.47No image inputGLM 5.1unstatedOpenwins 57%—$1.40 / $4.40Context: 1M → 203K+1ModelScoreCostPer taskWhat changes
Gemini 3.8 Flashhighwins 61%—*$1.24Sampling parameters are ignored+1Gemini 3 Flash Previewunstatedwins 61%—$0.50 / $3Gemini 3.1 Pro Previewunstatedwins 60%—$2 / $12Cannot use both thinking controlsGemini 2.5 Prounstatedwins 60%—$1.25 / $10Gemini 3.6 Flashhighwins 59%—*$1.60Gemini 3.5 Flash Liteunstatedwins 55%—$0.30 / $2.50Gemini 2.5 Flashunstatedsame—$0.30 / $2.50Gemma 3 12BunstatedOpenwins 43%—$0.05 / $0.10Context: 1M → 131KAll Google models →History
3 versionsBenchmarks
in Chat & KnowledgeAA Intelligence Index 4.3.2≈12.4#177
AA Omniscience Index-39.9#174
AA Omniscience · Accuracy26.8#97
Humanity's Last Exam8.7#208
GPQA Diamond76.6#138
IFBench (instructions)43.5#137
CritPt (physics)0#179
AIME 2025 · AA archived56.7#83
FAQ
How much does Gemini 2.5 Flash Preview (Sep '25) cost?›
There is no per-unit API price: not available via API.
Is Gemini 2.5 Flash Preview (Sep '25) free to use?›
There is no free API.
How good is Gemini 2.5 Flash Preview (Sep '25), and what is it best at?›
#207 of 334 in Chat & Knowledge, #110 of 239 in Coding, #151 of 242 in Agents & Tool Use, #85 of 195 in Writing & Language, #67 of 124 in Vision & Documents.
Is Gemini 2.5 Flash Preview (Sep '25) better than DeepSeek V4.1 Flash 1017?›
On Arena Text · German, Gemini 2.5 Flash Preview (Sep '25) scores 1,418 vs 1,418.6 — about the same.
What is the context window of Gemini 2.5 Flash Preview (Sep '25)?›
1M tokens.
When was Gemini 2.5 Flash Preview (Sep '25) released? Is there a newer version?›
Released 2025-09-25. It is the current version in its line.
Is Gemini 2.5 Flash Preview (Sep '25) open source?›
No, Gemini 2.5 Flash Preview (Sep '25) is a closed model available through an API. The closest open-weights option here is GLM 5.3.
Can Gemini 2.5 Flash Preview (Sep '25) read images?›
It accepts text and images, PDF files, audio, video as input.
What is Gemini 2.5 Flash Preview (Sep '25)?›
Gemini 2.5 Flash Preview (Sep '25) is a closed language model from Google, released Sep 2025.
What are the best alternatives to Gemini 2.5 Flash Preview (Sep '25)?›
GPT-5.6 Luna (better), GLM 5.3 (open weights).