Inference
EN
Google

Gemini 3.5 Flash-Lite

Gemini 3.5 Flash-Lite is a reasoning model from Google. 33 benchmarks count toward its index, in 8 categories.

  • Closed weights
  • Reasoning
  • Input: text · image · video · file · audio
  • Released 2026-07
RankedEnough results for an index.
IndexAll categories, out of 100.52.0±2.8Median 51.0
CoverageIndex weight that has results.100%
SpeedOutput tokens per second.130/sMedian 67/s
First tokenSeconds to the first output token.0.8 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.30Median $0.60
Output / 1MUS dollars per 1M output tokens.$2.50Median $2.55
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.1435±4 #90Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
google/gemini-3.5-flash-lite
Max outputMaximum output tokens per request.
66K
Knowledge cutoffLast date of training data.
—
ToolsTool calls on OpenRouter.
Yes
JSON outputOutput follows a given JSON schema.
Yes
Cache read / 1MUS dollars per 1M cached tokens.
$0.03
WeightsLink to the published weights.
—
RetiresDate OpenRouter removes the model.
—

Where to run

APIs that serve this model.
3 providers
ProviderTypeMaker, cloud, host or router.InputOutputContext
Google AI StudioMaker$0.30$2.501.05M
Google Vertex AIMaker$0.30$2.501.05M
OpenRouterRouter$0.30$2.501.05M

US dollars per 1M tokens, list price. From models.dev (opens in a new tab) (MIT).

Capabilities

Index per category, out of 100.
Out of 100
AgenticMulti-step tasks with tools.4 counted54.7105 / 218
CodingCode writing and repair.8 counted54.4102 / 209
ReasoningLogic problems and puzzles.8 counted47.7134 / 227
MultimodalTasks with images and text.2 counted58.845 / 84
KnowledgeFacts and expert knowledge.5 counted50.0135 / 268
MultilingualTasks in many languages.1 counted54.2—
InstructionTasks with strict prompt rules.1 counted62.6—
MathMath problems.4 counted46.599 / 168

Results

33 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
EuroEval SpanishMultilingual57.1% ±0.522 / 176
EuroEval PortugueseMultilingual61.3% ±0.525 / 182
EuroEval SwedishMultilingual63.0% ±0.627 / 195
EuroEval ItalianMultilingual61.6% ±0.525 / 174
EuroEval GermanMultilingual52.4% ±0.629 / 191
EuroEval DutchMultilingual59.6% ±0.527 / 175
EuroEval FrenchMultilingual61.6% ±0.335 / 178
EuroEval PolishMultilingual54.3% ±0.548 / 209
Artificial Analysis MMMU-ProMultimodal79.0%30 / 113
IOI v1Coding26.2% ±1.217 / 61
MMMU ProMultimodal83.6% ±0.930 / 92
GPQA diamond · Epoch AIKnowledge83.3% ±2.777 / 217
MMLU ProKnowledge85.8% ±0.353 / 135
GPQA Diamond · Vals AIKnowledge83.8% ±1.854 / 135
Artificial Analysis Long Context ReasoningReasoning76.0%88 / 217
Artificial Analysis GPQA DiamondKnowledge83.8%87 / 199
OSWorld-VerifiedAgentic74.0%18 / 37
OTIS Mock AIME 2024-2025Math71.1% ±6.898 / 200
SWE-benchCoding75.0% ±1.944 / 87
ARC-AGI-2 (semi-private)Reasoning10.3%67 / 131
LiveCodeBenchCoding79.0% ±1.174 / 142
LiveBench Instruction FollowingInstruction67.2%35 / 66
Artificial Analysis Humanity's Last ExamKnowledge18.8%118 / 222
MRCRv2Reasoning72.2%8 / 14
GDPval-AA normalizedAgentic24.2%76 / 133
Artificial Analysis Omniscience AccuracyKnowledge26.5%94 / 164
Mystery Game PuzzlesReasoning19.0% ±3.946 / 80
ARC-AGI-1 (semi-private)Reasoning53.5%73 / 126
Vibe Code Bench v1.1Coding37.2% ±4.678 / 123
LiveBench CodingCoding76.1%44 / 66
Critical Physics TasksReasoning0.0%150 / 214
SWE-bench ProCoding54.2%59 / 83
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal59.5%28 / 39
Artificial Analysis SciCodeCoding41.3%82 / 114
Terminal-Bench 2.1 · Vals AIAgentic50.2% ±1.058 / 75
Chess PuzzlesReasoning16.0% ±3.771 / 90
FrontierMath-Tiers-1-3-v2-PrivateMath26.0% ±2.672 / 90
LiveBench Agentic CodingAgentic45.3%53 / 66
Terminal-Bench 2.1 · BenchLMAgentic54.0%70 / 86
LiveBench LanguageKnowledge71.8%58 / 66
Code MigrationCoding6.1% ±1.770 / 77
FrontierMath-Tier-4-v2-PrivateMath0.0%70 / 72
LiveBench Data AnalysisReasoning53.2%78 / 80
LiveBench MathematicsMath73.7%65 / 66
LiveBench ReasoningReasoning60.2%79 / 80
MortgageTaxFinance68.7% ±0.911 / 97
LegalBenchLegal84.1% ±0.439 / 149
MedCodeHealthcare43.5% ±2.044 / 106
TaxEval v2Finance72.6% ±0.960 / 143
CorpFin v2Finance60.7% ±1.065 / 131
ARC-AGI-2 (public eval)Reasoning7.9%61 / 122
ProgramBenchCoding0.0%34 / 65
ARC-AGI-1 (public eval)Reasoning66.4%61 / 116
Finance Agent (v2)Finance47.4% ±0.553 / 79
CyberBench v1.1Other59.1% ±6.238 / 48
EMBFinance43.2% ±3.160 / 75
Legal Research BenchLegal13.9% ±2.476 / 90
MedScribeHealthcare70.9% ±2.093 / 109
Public Benefits Bench v1.1Public benefits51.8% ±1.344 / 50
Tax Agent BenchFinance7.3% ±1.973 / 82
SAGEEducation79.5% ±2.026 / 29
Artificial Analysis EnterpriseOps-GymAgentic42.3%—

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseVals AINo licence statedEpoch AICC BYLiveBenchNo licence statedEuroEvalMITARC Prize FoundationTerms restrict reusemodels.devMIT

History

5 releases
Index53.1 → 52.0
23 Oct 202612 Nov 2026
Output / 1M$2.50 → $2.50
23 Oct 202612 Nov 2026

Questions

How much does Gemini 3.5 Flash-Lite cost?
$0.30 per 1M input tokens, $2.50 per 1M output tokens.
How good is Gemini 3.5 Flash-Lite?
Its index is 52.0 out of 100. That is place 131 of 283 ranked models.
What is Gemini 3.5 Flash-Lite best at?
Agentic: 54.7 out of 100, place 105 of 218.
How fast is Gemini 3.5 Flash-Lite?
130 output tokens per second, 0.8 s to the first token.
What is the context window of Gemini 3.5 Flash-Lite?
1.05M tokens.
Is Gemini 3.5 Flash-Lite open source?
No. Its weights are not published.
When was Gemini 3.5 Flash-Lite released?
July 2026.

Badge

Gemini 3.5 Flash-Lite on Inference360: place 131 of 283, index 52.0
HTML
<a href="https://inference360.ai/models/gemini-3-5-flash-lite"><img src="https://inference360.ai/badges/gemini-3-5-flash-lite.svg" alt="Gemini 3.5 Flash-Lite on Inference360: place 131 of 283, index 52.0" height="20"></a>
Markdown
[![Gemini 3.5 Flash-Lite on Inference360: place 131 of 283, index 52.0](https://inference360.ai/badges/gemini-3-5-flash-lite.svg)](https://inference360.ai/models/gemini-3-5-flash-lite)

Compare with

Claude Opus 4.552.0Qwen3.6-27B52.1Solar Pro 452.1A.X K251.9dots3-note Preview63.4

Lower price

dots3-note Preview63.4 · FreeApodex 1.1 Mini55.7 · FreeNemotron 3 Ultra50.8 · Free

More from Google

All models
Gemini 3.6 Flash63.7Gemini 3.5 Flash CyberUnrankedNano Banana 2 Lite (Gemini 3.1 Flash Lite Image)UnrankedGemini 3.7 Flash68.5Nano Banana 2 (Gemini 3.1 Flash Image)Unranked