Inference
EN
Google

Gemini 3.6 Flash

Gemini 3.6 Flash is a reasoning model from Google. 38 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 100.63.7±2.6Median 51.0
CoverageIndex weight that has results.100%
SpeedOutput tokens per second.121/sMedian 67/s
First tokenSeconds to the first output token.1.6 sMedian 1.3 s
Input / 1MUS dollars per 1M input tokens.$0.75batch $0.375Median $0.60
Output / 1MUS dollars per 1M output tokens.$3.75batch $1.88 Batch price per 1M output tokens.Median $2.55
ContextMaximum tokens in one request.1.05MMedian 262K
EloLMArena rating and rank.1479±4 #22Median 1422

Details

From OpenRouter
API IDThe model ID on OpenRouter.
google/gemini-3.6-flash
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.
$0.075
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$0.75$3.751.05M
Google Vertex AIDeprecatedMaker$0.75$3.751.05M
OpenRouterRouter$0.75$3.751.05M

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.7 counted59.769 / 218
CodingCode writing and repair.9 counted62.058 / 209
ReasoningLogic problems and puzzles.8 counted62.159 / 227
MultimodalTasks with images and text.2 counted68.514 / 84
KnowledgeFacts and expert knowledge.6 counted67.929 / 268
MultilingualTasks in many languages.1 counted70.1—
InstructionTasks with strict prompt rules.1 counted79.1—
MathMath problems.4 counted60.747 / 168

Results

38 counted
BenchmarkThe benchmark name.ResultThe score from the source.PlacePlace among results on this benchmark.
EuroEval PolishMultilingual65.9% ±0.44 / 209
EuroEval SwedishMultilingual75.9% ±0.44 / 195
EuroEval GermanMultilingual66.0% ±0.54 / 191
EuroEval PortugueseMultilingual72.6% ±0.34 / 182
EuroEval FrenchMultilingual75.6% ±0.34 / 178
EuroEval SpanishMultilingual66.8% ±0.44 / 176
EuroEval DutchMultilingual71.8% ±0.34 / 175
EuroEval ItalianMultilingual75.1% ±0.34 / 174
GPQA Diamond · Vals AIKnowledge93.4% ±1.37 / 135
GPQA diamond · Epoch AIKnowledge94.1% ±1.412 / 217
LiveCodeBenchCoding88.1% ±0.98 / 142
Artificial Analysis GPQA DiamondKnowledge92.8%14 / 199
MMMU ProMultimodal88.4% ±0.87 / 92
MMLU ProKnowledge89.3% ±0.312 / 135
Artificial Analysis MMMU-ProMultimodal83.2%11 / 113
Artificial Analysis Omniscience AccuracyKnowledge50.0%20 / 164
LiveBench Instruction FollowingInstruction75.4%9 / 66
SimpleQA VerifiedKnowledge66.2% ±1.514 / 91
OSWorld-VerifiedAgentic83.0%6 / 37
Artificial Analysis Long Context ReasoningReasoning80.0%43 / 217
Artificial Analysis Humanity's Last ExamKnowledge40.8%44 / 222
OTIS Mock AIME 2024-2025Math94.2% ±3.148 / 200
LiveBench LanguageKnowledge83.9%16 / 66
ARC-AGI-1 (semi-private)Reasoning91.2%32 / 126
Terminal-Bench 2.1 · Vals AIAgentic73.8% ±1.621 / 75
SWE-benchCoding79.6% ±1.825 / 87
Critical Physics TasksReasoning10.6%63 / 214
ARC-AGI-2 (semi-private)Reasoning60.4%40 / 131
Artificial Analysis SciCodeCoding53.4%39 / 114
GDPval-AA normalizedAgentic39.2%46 / 133
ParseBench: document parsing for AI agents (LlamaIndex leaderboard)Multimodal69.1%14 / 39
Mystery Game PuzzlesReasoning30.0% ±4.630 / 80
Chess PuzzlesReasoning40.0% ±4.936 / 90
Vibe Code Bench v1.1Coding64.0% ±4.356 / 123
MazeBench: visual spatial reasoning in a 3D open worldAgentic0.0%23 / 48
FrontierMath-Tiers-1-3-v2-PrivateMath58.9% ±2.944 / 90
LiveBench ReasoningReasoning85.1%40 / 80
LiveBench CodingCoding77.9%33 / 66
Code MigrationCoding30.9% ±4.140 / 77
FrontierMath-Tier-4-v2-PrivateMath22.0% ±6.548 / 72
LiveBench MathematicsMath86.4%47 / 66
MirrorCodeCoding23.3% ±9.333 / 39
LiveBench Agentic CodingAgentic43.4%56 / 66
DeepSWEAgentic49.0%51 / 59
IOICoding35.1% ±5.739 / 45
LiveBench Data AnalysisReasoning63.0%70 / 80
cursorBench32Coding53.5%16 / 18
Furniture AssemblyReasoning23.3% ±5.433 / 37
Agent Arena command recoveryAgentic−18.246 / 50
LegalBenchLegal86.7% ±0.412 / 149
MedCodeHealthcare53.2% ±2.210 / 106
TaxEval v2Finance74.9% ±0.826 / 143
Finance Agent (v2)Finance56.3% ±0.219 / 79
ARC-AGI-1 (public eval)Reasoning96.0%28 / 116
ARC-AGI-2 (public eval)Reasoning65.3%30 / 122
MortgageTaxFinance67.6% ±0.924 / 97
EMBFinance65.4% ±2.625 / 75
CorpFin v2Finance63.3% ±0.944 / 131
ProgramBenchCoding0.0%34 / 65
MedScribeHealthcare79.7% ±1.961 / 109
Terminal-Bench-ScienceScience4.3% ±2.426 / 44
Tax Agent BenchFinance18.0% ±2.857 / 82
Legal Research BenchLegal25.0% ±3.065 / 90
Public Benefits Bench v1.1Public benefits56.6% ±1.339 / 50
CyberBench v1.1Other45.4% ±3.742 / 48
BioMysteryBenchScience58.5% ±1.629 / 30

Sources

BenchLMCC BY-NC 4.0OpenRouterTerms restrict reuseEpoch AICC BYVals AINo licence statedARC Prize FoundationTerms restrict reuseLiveBenchNo licence statedEuroEvalMITLMArenaCC BY 4.0models.devMIT

History

5 releases
Index65.2 → 63.7
23 Oct 202612 Nov 2026
Output / 1M$3.75 → $3.75
23 Oct 202612 Nov 2026

Questions

How much does Gemini 3.6 Flash cost?
$0.75 per 1M input tokens, $3.75 per 1M output tokens.
How good is Gemini 3.6 Flash?
Its index is 63.7 out of 100. That is place 55 of 283 ranked models.
What is Gemini 3.6 Flash best at?
Knowledge: 67.9 out of 100, place 29 of 268.
How fast is Gemini 3.6 Flash?
121 output tokens per second, 1.6 s to the first token.
What is the context window of Gemini 3.6 Flash?
1.05M tokens.
Is Gemini 3.6 Flash open source?
No. Its weights are not published.
When was Gemini 3.6 Flash released?
July 2026.

Badge

Gemini 3.6 Flash on Inference360: place 55 of 283, index 63.7
HTML
<a href="https://inference360.ai/models/gemini-3-6-flash"><img src="https://inference360.ai/badges/gemini-3-6-flash.svg" alt="Gemini 3.6 Flash on Inference360: place 55 of 283, index 63.7" height="20"></a>
Markdown
[![Gemini 3.6 Flash on Inference360: place 55 of 283, index 63.7](https://inference360.ai/badges/gemini-3-6-flash.svg)](https://inference360.ai/models/gemini-3-6-flash)

Compare with

Grok 4.563.7Qwen3.8 Max (0902)63.4Gemini 3.5 Flash-Lite52.0Gemma 4 12B39.0

Lower price

dots3-note Preview63.4 · FreeGLM-5.3-Flash63.1 · $0.078/1MLing 3.1 Flash67.2 · $0.135/1M

More from Google

All models
Gemini 3.5 Flash-Lite52.0Gemini 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