BenchLeader

Gemini 3.1 Flash Lite

Reasoning effort
Quality#135 of 333
52.5Index
Above the median
+2.5 on the median
Price#97 of 249
$0.563per 1M tokens
Cheaper than most
0.7× the median
Speed#1 of 215
359tokens/s
Top 10%
4.7× the median

29 cheaper ranked models score higher: DeepSeek V4.1 Flash 1017 (59.2, $0.525), DeepSeek V4.1 Flash (59.0, $0.518), GPT-5.6 Luna (58.5, $0.450) and 26 more.

As of 8 Nov 2026, Gemini 3.1 Flash Lite by Google scores 52.5 on the BenchLeader Index (593 ranked configurations), with its best results in long context (62) and weakest in reasoning (42). API pricing: $0.250 per 1M input tokens and $1.50 per 1M output tokens ($0.563 blended at 3:1), list price on 8 Nov 2026. It streams about 360 tokens/s (Artificial Analysis). Released 3 Mar 2026.

Key facts (as of 8 Nov 2026)

  • Gemini 3.1 Flash Lite API pricing: $0.250 per 1M input tokens and $1.50 per 1M output tokens ($0.563 blended at 3:1) (list price, 8 Nov 2026).
  • Gemini 3.1 Flash Lite streams about 360 tokens/s (Artificial Analysis, 8 Nov 2026).
  • Gemini 3.1 Flash Lite was released on 3 Mar 2026 (Google).

In detail. Gemini 3.1 Flash Lite is a Google proprietary reasoning model, released 3 Mar 2026. Its best configuration ranks #269 of 593 on the BenchLeader Index at 52.5 ±4.9, in the upper half. The ± is the point: 284 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 (62) and lowest in reasoning (42). At $0.563 per million tokens blended it is cheaper than most ranked models. Output speed of 359 tokens per second puts it in the fastest quarter, with a first answer in 13.6 s. It has been measured at 6 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 6 Nov 2026.

Blended price
$0.563/M
$0.250 in · $1.50 out
via Artificial Analysis
Output speed
359 tok/s
measured by Artificial Analysis
First answer
14 s
first token 0.78 s
Context
1.0M
via models.dev
Full answer
7.38 s
median, reasoning included
Cost per task
$0.194
44 s a task, AA Index
Tokens per task
16k
68% of it reasoning
Released
3 Mar 2026
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index53
  2. Reasoning42
  3. Coding43
  4. Agents & tools43
  5. Maths61
  6. Knowledge44
  7. Instruction following61
  8. Human preference61
  9. Multimodal58
  10. Long context62

Get an email when Google ships a new model

And once when it gets ranked, plus the weekly review on Mondays.

Choose providers and emails

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.

ModelReleasedIndexRank
Gemini 3.5 Flash Lite21 Jul 202651.7#297
Gemini 3.1 Flash Litethis page3 Mar 202652.5#269
Gemini 2.5 Flash Lite 09thinking25 Sep 202549.7#346
Gemini 2.5 Flash-Lite17 Jun 202548.3#377
Gemini 2.0 Flash Lite 00125 Feb 2025––
Gemini 2.0 Flash-Lite5 Feb 202548.3#375

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.

EffortIndexRankSpeedFirst answerChat reply costCategories
no reasoning––359 tok/s14 s$0.0006
minimal––359 tok/s14 s$0.0006Maths 38 · Reasoning 53
low––359 tok/s14 s$0.0006Maths 39 · Reasoning 54
medium––359 tok/s14 s$0.0006
high48.5#368359 tok/s14 s$0.0006Agents & tools 29 · Knowledge 55 · Maths 50 · Reasoning 59
not statedbest52.5#269359 tok/s14 s$0.0006Agents & tools 43 · Coding 43 · Human preference 61 · Instruction following 61 · Knowledge 44 · Long context 62 · Maths 61 · Multimodal 58 · Reasoning 42

How its index has moved

8 Sep 2026 to 8 Nov 2026
435260
  • 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

Benchmarkminimallowhighnot statedSourceTrend
GPQA Diamond73.7%#16774.2%#16481.8%#126–Epoch AI Benchmarking Hub
Humanity's Last Exameffort not stated–––8.6%#31Scale AI / CAIS
Arena Hard Promptseffort not stated–––1446#120Arena (LMArena)
GPQA Diamond (AA)not in indexeffort not stated–––82.2%#176Artificial Analysis
Humanity's Last Exam (AA)not in indexeffort not stated–––17.2%#247Artificial Analysis
HLE-Diamond (AA)effort not stated–––9.0%#110Artificial Analysis
GPQA Diamond (Vals)not in index––81.1%#65–Vals AI
Kagi LLM Benchmarkeffort not stated–––67.2%#34Kagi LLM Benchmark
CritPtnot in indexeffort not stated–––1.1%#234Artificial Analysis
CritPt v2 hard subset (AA)effort not stated–––3.0%#103Artificial Analysis
LMCAnot in index––35.0%#126–Epoch AI Benchmarking Hub
DTBenchnot in index––76.8%#146–Epoch AI Benchmarking Hub
ForecastBenchnot in indexeffort not stated–––62.8%#51ForecastBench

Coding

Benchmarkno reasoningmediumhighnot statedSourceTrend
SciCodeeffort not stated–––43.4%#154SciCode
WeirdMLeffort not stated–––52.2%#68WeirdML
Arena Codingeffort not stated–––1458#140Arena (LMArena)
Arena WebDeveffort not stated–––1256#132Arena (LMArena)
SciCode (AA)not in indexeffort not stated–––43.4%#163Artificial Analysis
LiveCodeBencheffort not stated–––80.1%#70Vals AI
SWE-bench (Vals)not in index––62.8%#74–Vals AI
Code Migrationnot in index––4.6%#73–Vals AI
ProgramBenchnot in index––0.0%#33–Vals AI
Vibe Code Bench v1.1not in index––0.0%#120–Vals AI
ALE-Benchnot in indexeffort not stated–––797.7#79Epoch AI Benchmarking Hub
Terminal-Bench 4.0 (AA)not in indexeffort not stated–––0.5%#157Artificial Analysis
Terminal-Bench 2.1 (AA)not in indexeffort not stated–––31.1%#155Artificial Analysis
GameBench 2 by reasoning settingnot in index10.7#18619.2#18015.7#184–DuelLab GameBench 2
Design Arena Websitenot in indexeffort not stated–––1080#147Design Arena
FrontierCode 2.0 (AA)effort not stated–––0.6%#104Artificial Analysis

Agents & tools

Benchmarklowhighnot statedSourceTrend
Terminal-Bench Hardeffort not stated––24.2%#147Artificial Analysis
τ²-Bench Telecom (AA)not in indexeffort not stated––31.3%#258Artificial Analysis
Terminal-Bench 2.1 (Vals)–34.1%#72–Vals AI
MCP Atlas–57.1%#33–Scale AI SEAL
τ³-Banking (AA)effort not stated––9.7%#149Artificial Analysis
Analyst Agent (AA)not in indexeffort not stated––8.8%#48Artificial Analysis
APEX-Agents (AA)not in indexeffort not stated––12.2%#27Artificial Analysis
Finance Agent v2not in index–30.0%#73–Vals AI
Harvey's Legal Agent Benchmark v1not in index–0.0%#64–Vals AI
Legal Research Benchnot in index–3.4%#85–Vals AI
Tax Agent Benchnot in index–3.1%#78–Vals AI
Terminal-Bench 2.0 (Vals)not in index–24.7%#52–Vals AI
DeepResearch Benchnot in index37.3%#36–36.4%#37DeepResearch Bench
AutomationBenchnot in indexeffort not stated––6.8%#177Artificial Analysis
GDP.pdf (AA)not in indexeffort not stated––7.8%#164Artificial Analysis
EnterpriseOps-Gymnot in indexeffort not stated––28.0%#47Artificial Analysis
Terminal-Bench-Science 0.1 (AA)not in indexeffort not stated––0.0%#100Artificial Analysis
BrowseCompnot in indexeffort not stated––1.1%#17OpenAI

Maths

Benchmarkminimallowhighnot statedSourceTrend
FrontierMath Tiers 1–321.4%#8922.5%#8727.7%#80–Epoch AI Benchmarking Hub
OTIS Mock AIME37.8%#21044.4%#20380.0%#123–Epoch AI Benchmarking Hub
AIME (Vals)––83.3%#43–Vals AI
Arena Mathseffort not stated–––1436#101Arena (LMArena)

Knowledge

Benchmarkhighnot statedSourceTrend
AA-Omniscience v1.1effort not stated–-24.1#113Artificial Analysis
MMLU-Proeffort not stated–86.2%#44Vals AI
LegalBench83.8%#48–Vals AI
CorpFin59.4%#78–Vals AI
TaxEval71.8%#72–Vals AI
MultiNRCeffort not statedpossible contamination–25.0%#29Scale AI SEAL
CaseLaw v2not in index55.0%#35–Vals AI
Excel Modeling Benchmarknot in index8.6%#76–Vals AI
MedCodenot in index47.6%#32–Vals AI
MedScribenot in index63.9%#101–Vals AI
AA-Omniscience v1.1: accuracynot in indexeffort not stated–31.7%#84Artificial Analysis
AA-Omniscience v1.1: non-hallucinationnot in indexeffort not stated–18.3%#101Artificial Analysis

Instruction following

Human preference

Benchmarknot statedSourceTrend
Arena Text1433#110Arena (LMArena)
Arena Creative Writingnot in index1414#89Arena (LMArena)
Arena Multi-turnnot in index1435#109Arena (LMArena)
Arena Longer Queriesnot in index1430#126Arena (LMArena)

Multimodal

Benchmarkhighnot statedSourceTrend
Arena Visioneffort not stated–1240#72Arena (LMArena)
MMMU-Proeffort not stated–75.5%#107Artificial Analysis
VISTAeffort not stated–46.9%#19Scale AI SEAL
MMMU-Pro (Vals)not in index82.5%#37–Vals AI
MortgageTaxnot in index68.0%#20–Vals AI

Long context

Benchmarknot statedSourceTrend
AA-LCR v1.174.3%#171Artificial Analysis
MLCRnot in index5.0%#149Artificial Analysis

Composite

Benchmarkhighnot statedSourceTrend
Epoch Capabilities Indexnot in indexeffort not stated–144.9#88Epoch AI Benchmarking Hub
AA Intelligence Index v5.0not in indexeffort not stated–6.2#101Artificial Analysis
AA Intelligence Index v4.3.2not in indexeffort not stated–15.6#290Artificial Analysis
AA Progress Index v1not in indexeffort not stated–244.1#280Artificial Analysis
Vals Multimodal Indexnot in index41.4%#31–Vals AI

Safety & honesty

Benchmarknot statedSourceTrend
FORTRESSnot in index51.0%#14Scale AI SEAL
MASKnot in indexpossible contamination48.4%#57Scale AI SEAL

GameBench 2, best reasoning setting (Medium): 19.2. Reference only, not part of the BenchLeader Index.

DuelLab GameBench 2, release gb2-20261103-eaea31df94e6 (https://benchmarks.duellab.org/releases/gb2-20261103-eaea31df94e6/), CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Scores mapped to BenchLeader model names; not endorsed by DuelLab.

API pricing: 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.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Google AI Studio170 tok/s0.58 s$0.450$2.701.0M–
Google AI Studio150 tok/s0.73 s$0.250$1.501.0M–
Google Vertex99 tok/s1.14 s$0.450$2.701.0M–
Google AI Studio97 tok/s0.61 s$0.125$0.7501.0M–
Google Vertex86 tok/s1.29 s$0.250$1.501.0M–
Google Vertex33 tok/s6.02 s$0.275$1.651.0M–
Google Vertex14 tok/s41 s$0.125$0.7501.0M–

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 $0.025 per 1M. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0006$0.000514.4 s
Summarize a 30-page report12,000 / 600$0.0039$0.001915.3 s
Code edit6,000 / 1,500$0.0038$0.002717.8 s
Agentic coding session60,000 / 4,000$0.021$0.01124.7 s
Structured extraction2,000 / 200$0.0008$0.000514.1 s

More from Google

New this month

See also

Embed or cite

Show where Gemini 3.1 Flash Lite ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page. To see ranks like this on OpenRouter and Hugging Face model pages, get the Chrome extension.

[![Gemini 3.1 Flash Lite on BenchLeader: #134 overall](https://www.benchleader.com/badge/gemini-3-1-flash-lite.svg)](https://www.benchleader.com/models/gemini-3-1-flash-lite)

Cite this

The ranking changes daily, so a reference carries the data date. APA style:

BenchLeader. (2026, November 8). Gemini 3.1 Flash Lite on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 8, 2026, from https://www.benchleader.com/models/gemini-3-1-flash-lite

BibTeX:

@misc{benchleader_gemini_3_1_flash_lite_20261108,
  author       = {{BenchLeader}},
  title        = {{Gemini 3.1 Flash Lite on the BenchLeader Index (method v1.4)}},
  year         = {2026},
  month        = nov,
  howpublished = {\url{https://www.benchleader.com/models/gemini-3-1-flash-lite}},
  note         = {Data as of 2026-11-08}
}

Every badge style, the embeddable top-10 widget and how to cite the Index

Questions people ask

What are Gemini 3.1 Flash Lite's benchmark scores?
As of 8 Nov 2026, Gemini 3.1 Flash Lite scores 52.5 ±4.9 on the BenchLeader Index, strongest in long context (62) and weakest in reasoning (42). That rests on independent benchmark results only, and every score on this page links to its publisher.
How does Gemini 3.1 Flash Lite rank?
Gemini 3.1 Flash Lite is #269 of 593 ranked configurations on 8 Nov 2026. 284 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 69.4.
How much does Gemini 3.1 Flash Lite cost per 1M tokens?
As of 8 Nov 2026, Gemini 3.1 Flash Lite costs $0.250 per 1M input tokens and $1.50 per 1M output tokens, $0.563 blended (three input tokens per output token).
When was Gemini 3.1 Flash Lite released?
Google released Gemini 3.1 Flash Lite on 3 Mar 2026. BenchLeader has results for it at 6 reasoning-effort settings.

Data as of 8 Nov 2026. Compare these configurations.

Cite as: BenchLeader, “Gemini 3.1 Flash Lite: benchmarks, pricing, speed and rank”, data as of 8 Nov 2026, https://www.benchleader.com/models/gemini-3-1-flash-lite.