GoogleGemini 2.5 Flash 09 2025
As of 29 Oct 2026, Gemini 2.5 Flash 09 2025 by Google scores 52.9 on the BenchLeader Index (581 ranked configurations), with its best results in long context (60) and weakest in agents & tools (48). Released 25 Sept 2025; 12 independent results so far.
Key facts (as of 29 Oct 2026)
- Gemini 2.5 Flash 09 2025 was released on 25 Sept 2025 (Google).
In detail. Gemini 2.5 Flash 09 2025 is a Google proprietary reasoning model, released 25 Sept 2025. Its best configuration (thinking reasoning effort) ranks #261 of 581 on the BenchLeader Index at 52.9 ±2.2, in the upper half. The ± is the point: 260 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in long context (60) and lowest in agents & tools (48). It has been measured at 3 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 13 Oct 2026.
- Blended price
- –
- Output speed
- –
- First answer
- –
- Context
- 1M
- Full answer
- –
- Cost per run
- –
- Released
- 25 Sept 2025
- Overall index53
- Coding55
- Agents & tools48
- Maths48
- Knowledge55
- Instruction following53
- Multimodal56
- Long context60
Versions
Google has shipped 12 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.
| Model | Released | Index | Rank |
|---|---|---|---|
| Gemini 3.8 Flashhigh | 2 Sept 2026 | 61.8 | #56 |
| Gemini 3.7 Flashhigh | 13 Aug 2026 | 62.2 | #49 |
| Gemini 3.6 Flashhigh | 21 Jul 2026 | 61.3 | #62 |
| Gemini 3.5 Flashhigh | 19 May 2026 | 62.6 | #42 |
| Gemini 3 Flashthinking | 17 Dec 2025 | 59.0 | #103 |
| Gemini 2.5 Flash 09 2025thinkingthis page | 25 Sept 2025 | 52.9 | #261 |
| Gemini 2.5 Flash | 17 Apr 2025 | 52.7 | #265 |
| Gemini 2.0 Flash | 5 Feb 2025 | 43.5 | #461 |
| Gemini 1.5 Flash 002 | 24 Sept 2024 | 38.3 | #537 |
| Gemini 1.5 Flash 001 | 23 May 2024 | – | – |
| Gemini 1.5 Flash | 14 May 2024 | 45.2 | #438 |
| Gemini Flash 2.0 | – | – | – |
Reasoning-effort configurations
The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.
| Effort | Index | Rank | Speed | First answer | Chat reply cost | Categories |
|---|---|---|---|---|---|---|
| no reasoning | – | – | – | – | – | Agents & tools 46 · Instruction following 46 · Long context 54 · Multimodal 53 |
| thinkingbest | 52.9 | #261 | – | – | – | Agents & tools 48 · Coding 55 · Instruction following 53 · Knowledge 55 · Long context 60 · Maths 48 · Multimodal 56 |
| not stated | 46.8 | #405 | – | – | – | Agents & tools 30 · Coding 51 · Knowledge 55 · Maths 48 |
How its index has moved
10 Sept 2026 to 29 Oct 2026- thinking
- not stated
The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.
Benchmark results
See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.
Metadata
| Benchmark | thinking | Source |
|---|---|---|
| AA Openness Indexnot in index | 5.6#51 | Artificial Analysis |
Reasoning
| Benchmark | no reasoning | thinking | not stated | Source |
|---|---|---|---|---|
| GPQA Diamond (AA)not in index | 76.6%#235 | 79.3%#202 | – | Artificial Analysis |
| Humanity's Last Exam (AA)not in index | 8.7%#355 | 13.8%#276 | – | Artificial Analysis |
| GPQA Diamond (Vals)not in index | – | 76.5%#75 | 81.6%#63 | Vals AI |
| CritPtnot in index | 0.0%#328 | 0.3%#283 | – | Artificial Analysis |
Coding
| Benchmark | no reasoning | thinking | not stated | Source |
|---|---|---|---|---|
| WeirdMLeffort not stated | – | – | 41.9%#94 | WeirdML |
| LiveCodeBench | – | 76.2%#77 | 75.1%#78 | Vals AI |
| LiveCodeBench (AA)not in index | 62.5%#103 | 71.3%#59 | – | Artificial Analysis |
Agents & tools
| Benchmark | no reasoning | thinking | not stated | Source |
|---|---|---|---|---|
| Terminal-Bencheffort not stated | – | – | 17.1%#75 | Terminal-Bench |
| Terminal-Bench Hard | 14.4%#198 | 16.7%#186 | – | Artificial Analysis |
| τ²-Bench Telecom (AA)not in index | 28.4%#269 | 45.6%#211 | – | Artificial Analysis |
| Terminal-Bench 2.0 (Vals)not in index | – | 21.4%#55 | – | Vals AI |
| TheAgentCompanynot in indexeffort not stated | – | – | 41.1%#2 | TheAgentCompany |
Maths
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| AIME (Vals) | 51.5%#58 | 49.8%#59 | Vals AI |
| MGSM | 89.8%#40 | 89.8%#39 | Vals AI |
Knowledge
| Benchmark | thinking | not stated | Source |
|---|---|---|---|
| MMLU-Pro | 83.7%#70 | 83.7%#69 | Vals AI |
| LegalBench | 82.6%#67 | 82.5%#70 | Vals AI |
| CorpFin | 59.8%#72 | 59.0%#78 | Vals AI |
| TaxEval | 72.4%#62 | 72.7%#58 | Vals AI |
| MedQA | 91.2%#40 | 91.4%#37 | Vals AI |
| MedCodenot in index | 40.3%#66 | 40.5%#64 | Vals AI |
| MedScribenot in index | 78.5%#65 | 78.0%#68 | Vals AI |
| Global-MMLU-Lite (AA)not in index | 88.8%#18 | – | Artificial Analysis |
Instruction following
| Benchmark | no reasoning | thinking | Source |
|---|---|---|---|
| IFBench | 43.5%#224 | 52.3%#163 | Artificial Analysis |
Multimodal
| Benchmark | no reasoning | thinking | not stated | Source |
|---|---|---|---|---|
| MMMU-Pro | 70.2%#154 | 73.1%#133 | – | Artificial Analysis |
| MMMU-Pro (Vals)not in index | – | 80.8%#43 | 79.5%#46 | Vals AI |
| MortgageTaxnot in index | – | 61.9%#57 | 62.6%#52 | Vals AI |
| VPCTnot in indexeffort not stated | – | – | 46.2%#12 | Epoch AI Benchmarking Hub |
Long context
| Benchmark | no reasoning | thinking | Source |
|---|---|---|---|
| AA-LCR v1.1 | 60.0%#292 | 71.0%#214 | Artificial Analysis |
Composite
| Benchmark | no reasoning | thinking | Source |
|---|---|---|---|
| AA Intelligence Index v4.3.2not in index | 12.4#344 | 15.5#292 | Artificial Analysis |
| AA Progress Index v1not in index | 217.1#351 | 235.3#296 | Artificial Analysis |
More from Google
See also
Questions people ask
- What are Gemini 2.5 Flash 09 2025's benchmark scores?
- As of 29 Oct 2026, Gemini 2.5 Flash 09 2025 (thinking effort) scores 52.9 ±2.2 on the BenchLeader Index, strongest in long context (60) and weakest in agents & tools (48). That rests on 12 independent benchmark results, and every score on this page links to its publisher.
- How does Gemini 2.5 Flash 09 2025 rank?
- Gemini 2.5 Flash 09 2025 (thinking effort) is #261 of 581 ranked configurations on 29 Oct 2026. 260 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 69.1.
- When was Gemini 2.5 Flash 09 2025 released?
- Google released Gemini 2.5 Flash 09 2025 on 25 Sept 2025. BenchLeader has results for it at 3 reasoning-effort settings.
Data as of 29 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Gemini 2.5 Flash 09 2025: benchmarks, pricing, speed and rank”, data as of 29 Oct 2026, https://www.benchleader.com/models/gemini-2-5-flash-09-2025.