GoogleGemini 3.5 Flash Lite
Gemini 3.5 Flash Lite is a Google proprietary reasoning model, released 21 Jul 2026. Its best configuration ranks #252 of 785 on the BenchLeader Index at 54.5 ±5.9, in the upper half. The ± is the point: 331 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 human preference (64) and lowest in agents & tools (40). At $0.850 per million tokens blended it is mid-priced. Output speed of 430 tokens per second puts it in the fastest quarter, with a first answer in 6.9 s. It has been measured at 5 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 22 Oct 2026.
- Blended price
- $0.850/M
- $0.300 in · $2.50 out
- Output speed
- 430 tok/s
- measured by Artificial Analysis
- First answer
- 6.94 s
- first token 0.62 s
- Context
- 1.0M
- Full answer
- 8.10 s
- median, reasoning included
- Cost per run
- $0.193
- one full Intelligence Index run
- Released
- 21 Jul 2026
- Overall index55
- Reasoning49
- Coding50
- Agents & tools40
- Maths54
- Knowledge64
- Human preference64
- Multimodal61
- Long context63
- Composite54
Versions
Google has shipped 6 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 3.5 Flash Litethis page | 21 Jul 2026 | 54.5 | #252 |
| Gemini 3.1 Flash Lite | 3 Mar 2026 | 53.7 | #274 |
| Gemini 2.5 Flash Lite 09thinking | 25 Sept 2025 | 47.9 | #444 |
| Gemini 2.5 Flash-Lite | 17 Jun 2025 | 48.3 | #433 |
| Gemini 2.0 Flash Lite 001 | 25 Feb 2025 | – | – |
| Gemini 2.0 Flash-Lite | 5 Feb 2025 | 47.2 | #462 |
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 |
|---|---|---|---|---|---|---|
| minimal | – | – | 430 tok/s | 6.94 s | $0.0009 | Maths 46 · Reasoning 40 |
| low | – | – | 430 tok/s | 6.94 s | $0.0009 | Maths 50 · Reasoning 41 |
| medium | – | – | 430 tok/s | 6.94 s | $0.0009 | Reasoning 35 |
| high | 47.9 | #445 | 430 tok/s | 6.94 s | $0.0009 | Agents & tools 40 · Coding 52 · Knowledge 57 · Maths 41 · Reasoning 48 |
| not statedbest | 54.5 | #252 | 430 tok/s | 6.94 s | $0.0009 | Agents & tools 40 · Coding 50 · Composite 54 · Human preference 64 · Knowledge 64 · Long context 63 · Maths 54 · Multimodal 61 · Reasoning 49 |
How its index has moved
8 Sept 2026 to 21 Oct 2026- not stated
- high
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.
Reasoning
| Benchmark | minimal | low | medium | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|---|
| GPQA Diamond | 74.2%#154 | 75.8%#144 | – | 83.3%#103 | – | Epoch AI Benchmarking Hub | |
| Arena Hard Promptseffort not stated | – | – | – | – | 1475#76 | Arena (LMArena) | |
| LiveBench 1.0 Reasoningnot in index | – | – | – | 60.2%#66 | – | LiveBench | |
| GPQA Diamond (AA)not in indexeffort not stated | – | – | – | – | 83.8%#153 | Artificial Analysis | |
| Humanity's Last Exam (AA)not in indexeffort not stated | – | – | – | – | 18.8%#226 | Artificial Analysis | |
| GPQA Diamond (Vals)not in index | – | – | – | 83.8%#54 | – | Vals AI | |
| ARC-AGI-1 | 7.5%#238 | 17.0%#224 | 32.3%#207 | 53.5%#176 | – | ARC Prize | |
| ARC-AGI-2 | 0.8%#231 | 1.5%#214 | 5.3%#174 | 10.3%#159 | – | ARC Prize | |
| CritPteffort not stated | – | – | – | – | 0.0%#315 | Artificial Analysis | |
| Chess Puzzlesnot in index | 7.0%#127 | 21.0%#88 | – | 16.0%#104 | – | Epoch AI Benchmarking Hub | |
| Mystery Game Puzzlesnot in index | 12.0%#103 | 19.0%#64 | – | – | – | Epoch AI Benchmarking Hub | |
| LMCAnot in index | – | – | – | 37.6%#116 | – | Epoch AI Benchmarking Hub | |
| DTBenchnot in index | – | – | – | 83.5%#113 | – | Epoch AI Benchmarking Hub |
Coding
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| SciCodeeffort not stated | – | 41.3%#154 | SciCode | |
| WeirdML | 39.0%#110 | – | WeirdML | |
| Arena Codingeffort not stated | – | 1502#76 | Arena (LMArena) | |
| Arena WebDeveffort not stated | – | 1434#81 | Arena (LMArena) | |
| LiveBench 1.0 Codingnot in index | 76.1%#47 | – | LiveBench | |
| SciCode (AA)not in indexeffort not stated | – | 41.3%#167 | Artificial Analysis | |
| LiveCodeBench | 79.0%#72 | – | Vals AI | |
| SWE-bench (Vals)not in index | 75.0%#42 | – | Vals AI | |
| Code Migrationnot in index | 6.2%#67 | – | Vals AI | |
| ProgramBenchnot in index | 0.0%#32 | – | Vals AI | |
| Vibe Code Bench v1.1not in index | 37.2%#70 | – | Vals AI | |
| ALE-Benchnot in index | 765.3#80 | – | Epoch AI Benchmarking Hub | |
| Terminal-Bench 4.0 (AA)not in indexeffort not stated | – | 1.0%#128 | Artificial Analysis | |
| Terminal-Bench 2.1 (AA)not in indexeffort not stated | – | 53.6%#111 | Artificial Analysis |
Agents & tools
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| APEX-Agentseffort not stated | – | 29.3%#49 | Mercor | |
| LiveBench Agentic Codingnot in index | 45.3%#54 | – | LiveBench | |
| Terminal-Bench 2.1 (Vals) | 50.2%#57 | – | Vals AI | |
| GDPval-AA v2.1not in indexeffort not stated | – | 24.2%#95 | Artificial Analysis | |
| τ³-Banking (AA)effort not stated | – | 17.5%#106 | Artificial Analysis | |
| CyberBenchnot in index | 59.1%#35 | – | Vals AI | |
| Finance Agent v2not in index | 47.4%#53 | – | Vals AI | |
| Harvey's Legal Agent Benchmark v1not in index | 0.0%#61 | – | Vals AI | |
| Legal Research Benchnot in index | 13.9%#68 | – | Vals AI | |
| Tax Agent Benchnot in index | 7.3%#62 | – | Vals AI | |
| AutomationBenchnot in indexeffort not stated | – | 25.0%#126 | Artificial Analysis | |
| GDP.pdf (AA)not in indexeffort not stated | – | 13.6%#104 | Artificial Analysis | |
| Harvey LAB-AA v1.1not in indexeffort not stated | – | 0.0%#23 | Harvey | |
| EnterpriseOps-Gymnot in indexeffort not stated | – | 42.4%#23 | Artificial Analysis | |
| Terminal-Bench-Science 0.1 (AA)not in indexeffort not stated | – | 0.0%#71 | Artificial Analysis | |
| FACTS Searchnot in indexeffort not stated | – | 63.4%#20 |
Maths
| Benchmark | minimal | low | high | not stated | Source | Trend |
|---|---|---|---|---|---|---|
| FrontierMath Tiers 1–3 | – | – | 26.0%#79 | – | Epoch AI Benchmarking Hub | |
| FrontierMath Tier 4 | – | – | 0.0%#72 | – | Epoch AI Benchmarking Hub | |
| OTIS Mock AIME | 51.1%#183 | 60.0%#163 | 71.1%#135 | – | Epoch AI Benchmarking Hub | |
| ProofBencheffort not stated | – | – | – | 13.0%#62 | Vals AI | |
| LiveBench Mathematicsnot in index | – | – | 73.7%#66 | – | LiveBench | |
| Arena Mathseffort not stated | – | – | – | 1443#88 | Arena (LMArena) |
Knowledge
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| LiveBench 1.0 Data Analysisnot in index | 53.3%#65 | – | LiveBench | |
| AA-Omniscienceeffort not stated | – | 5.2#107 | Artificial Analysis | |
| MMLU-Pro | 85.8%#51 | – | Vals AI | |
| LegalBench | 84.1%#40 | – | Vals AI | |
| CorpFin | 60.7%#65 | – | Vals AI | |
| TaxEval | 72.6%#59 | – | Vals AI | |
| Excel Modeling Benchmarknot in index | 43.2%#58 | – | Vals AI | |
| MedCodenot in index | 43.5%#44 | – | Vals AI | |
| MedScribenot in index | 70.9%#89 | – | Vals AI | |
| Public Benefits Bench v1.1not in index | 51.8%#42 | – | Vals AI | |
| GDP.pdfnot in index | 10.0%#50 | – | Epoch AI Benchmarking Hub | |
| AA-Omniscience: accuracynot in indexeffort not stated | – | 29.5%#188 | Artificial Analysis | |
| AA-Omniscience: non-hallucinationnot in indexeffort not stated | – | 65.6%#62 | Artificial Analysis |
Instruction following
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| LiveBench Languagenot in index | 71.8%#58 | – | LiveBench | |
| LiveBench Instruction Followingnot in index | 67.2%#36 | – | LiveBench | |
| Arena Instruction Followingnot in indexeffort not stated | – | 1443#75 | Arena (LMArena) |
Human preference
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| Arena Text | 1455#71 | Arena (LMArena) | |
| Arena Creative Writingnot in index | 1437#61 | Arena (LMArena) | |
| Arena Multi-turnnot in index | 1467#67 | Arena (LMArena) | |
| Arena Longer Queriesnot in index | 1462#71 | Arena (LMArena) |
Multimodal
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Arena Visioneffort not stated | – | 1269#45 | Arena (LMArena) | |
| MMMU-Proeffort not stated | – | 79.0%#65 | Artificial Analysis | |
| MMMU-Pro (Vals)not in index | 83.6%#30 | – | Vals AI | |
| MortgageTaxnot in index | 68.7%#12 | – | Vals AI | |
| SAGEnot in index | 47.3%#32 | – | Vals AI |
Long context
| Benchmark | not stated | Source | Trend |
|---|---|---|---|
| AA-LCR | 76.0%#148 | Artificial Analysis | |
| MLCRnot in index | 10.6%#108 | Artificial Analysis |
Composite
| Benchmark | high | not stated | Source | Trend |
|---|---|---|---|---|
| Epoch Capabilities Indexnot in indexeffort not stated | – | 144.9#86 | Epoch AI Benchmarking Hub | |
| LiveBench 1.0not in index | 63.9%#65 | – | LiveBench | |
| AA Intelligence Index v4.3.2effort not stated | – | 22.2#204 | Artificial Analysis |
Where to run it
Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.
| Provider | Speed | First token | Input $/M | Output $/M | Context | Quantisation |
|---|---|---|---|---|---|---|
| Google AI Studio | 319 tok/s | 1.03 s | $0.540 | $4.50 | 1.0M | – |
| Google Vertex | 272 tok/s | 0.99 s | $0.540 | $4.50 | 1.0M | – |
| Google Vertex | 168 tok/s | 0.96 s | $0.330 | $2.75 | 1.0M | – |
| Google AI Studio | 161 tok/s | 0.98 s | $0.150 | $1.25 | 1.0M | – |
| Google Vertex | 106 tok/s | 0.62 s | $0.300 | $2.50 | 1.0M | – |
| Google Vertex | 104 tok/s | 0.38 s | $0.330 | $2.75 | 1.0M | – |
| Google AI Studio | 81 tok/s | 0.63 s | $0.300 | $2.50 | 1.0M | – |
| Google Vertex | 33 tok/s | 6.74 s | $0.150 | $1.25 | 1.0M | – |
What a task costs
Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.030 per 1M. Reasoning tokens are not modelled.
| Workload | Tokens in / out | Cost | With caching | Time |
|---|---|---|---|---|
| Chat reply | 400 / 300 | $0.0009 | $0.0008 | 7.6 s |
| Summarize a 30-page report | 12,000 / 600 | $0.0051 | $0.0027 | 8.3 s |
| Code edit | 6,000 / 1,500 | $0.0056 | $0.0043 | 10.4 s |
| Agentic coding session | 60,000 / 4,000 | $0.028 | $0.016 | 16.2 s |
| Structured extraction | 2,000 / 200 | $0.0011 | $0.0007 | 7.4 s |
More from Google
See also
Questions people ask
- What are Gemini 3.5 Flash Lite's benchmark scores?
- As of 24 Oct 2026, Gemini 3.5 Flash Lite scores 54.5 ±5.9 on the BenchLeader Index, strongest in human preference (64) and weakest in agents & tools (40). That rests on 15 independent benchmark results, and every score on this page links to its publisher.
- How does Gemini 3.5 Flash Lite rank?
- Gemini 3.5 Flash Lite is #252 of 785 ranked configurations on 24 Oct 2026. 331 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 72.6.
- How much does Gemini 3.5 Flash Lite cost per 1M tokens?
- As of 24 Oct 2026, Gemini 3.5 Flash Lite costs $0.300 per 1M input tokens and $2.50 per 1M output tokens, $0.850 blended (three input tokens per output token).
- When was Gemini 3.5 Flash Lite released?
- Google released Gemini 3.5 Flash Lite on 21 Jul 2026. BenchLeader has results for it at 5 reasoning-effort settings.
Data as of 24 Oct 2026. Compare these configurations.
Cite as: BenchLeader, “Gemini 3.5 Flash Lite: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/gemini-3-5-flash-lite.