What are Gemini 3.6 Flash's benchmark scores?

As of 24 Oct 2026, Gemini 3.6 Flash (high effort) scores 62.0 ±2.9 on the BenchLeader Index, top 100, strongest in composite (68) and weakest in coding (57). That rests on 28 independent benchmark results, and every score on this page links to its publisher.

How does Gemini 3.6 Flash rank?

Gemini 3.6 Flash (high effort) is #89 of 785 ranked configurations on 24 Oct 2026 (top 100). 170 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 3.6 Flash cost per 1M tokens?

As of 24 Oct 2026, Gemini 3.6 Flash costs $0.750 per 1M input tokens and $3.75 per 1M output tokens, $1.50 blended (three input tokens per output token). That is a launch price until 31 Dec 2026, then $3.00 blended.

Is Gemini 3.6 Flash better than GPT-5.6 Sol?

As of 24 Oct 2026, GPT-5.6 Sol (max) leads on quality (68.0 vs 62.0 on the BenchLeader Index), but Gemini 3.6 Flash (high) costs 5.3× less, so it is the better buy unless you need the extra quality. The head-to-head page has the benchmark-by-benchmark detail.

When was Gemini 3.6 Flash released?

Google released Gemini 3.6 Flash on 21 Jul 2026. BenchLeader has results for it at 5 reasoning-effort settings.

GoogleReasoning modelCompare with another model

Gemini 3.6 Flash

Reasoning effort

Gemini 3.6 Flash is a Google proprietary reasoning model, released 21 Jul 2026. Its best configuration (high reasoning effort) ranks #89 of 785 on the BenchLeader Index at 62.0 ±2.9, in the upper half. The ± is the point: 170 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 composite (68) and lowest in coding (57). At $1.50 per million tokens blended it is pricier than most ranked models; that price runs until 31 Dec 2026, then $3.00. Output speed of 202 tokens per second puts it in the fastest quarter, with a first answer in 15.8 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: $1.50/M; $0.750 in · $3.75 out
Output speed: 202 tok/s; measured by Artificial Analysis
First answer: 16 s; first token 1.35 s
Context: 1.0M
Full answer: 18 s; median, reasoning included
Cost per run: $1.60; one full Intelligence Index run
Released: 21 Jul 2026

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index62
Reasoning63
Coding57
Agents & tools59
Maths60
Knowledge66
Human preference68
Multimodal65
Long context65
Composite68

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	64.4	#50
Gemini 3.7 Flashhigh	13 Aug 2026	63.5	#64
Gemini 3.6 Flashhighthis page	21 Jul 2026	62.0	#89
Gemini 3.5 Flashhigh	19 May 2026	63.1	#70
Gemini 3 Flashthinking	17 Dec 2025	59.1	#138
Gemini 2.5 Flash 09 2025thinking	25 Sept 2025	51.1	#351
Gemini 2.5 Flash	17 Apr 2025	52.7	#306
Gemini 2.0 Flash	5 Feb 2025	43.2	#574
Gemini 1.5 Flash 002	24 Sept 2024	38.3	#716
Gemini 1.5 Flash 001	23 May 2024	–	–
Gemini 1.5 Flash	14 May 2024	44.3	#541
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
minimal	–	–	123 tok/s	1.35 s	$0.0014	Maths 58 · Reasoning 46
low	–	–	123 tok/s	1.35 s	$0.0014	Maths 59 · Reasoning 55
medium	–	–	123 tok/s	1.35 s	$0.0014	Reasoning 57
highbest	62.0	#89	202 tok/s	16 s	$0.0014	Agents & tools 59 · Coding 57 · Composite 68 · Human preference 68 · Knowledge 66 · Long context 65 · Maths 60 · Multimodal 65 · Reasoning 63
not stated	49.5	#395	123 tok/s	1.35 s	$0.0014	Agents & tools 47 · Coding 47 · Maths 54

How its index has moved

8 Sept 2026 to 23 Oct 2026

high
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.

Reasoning

Benchmark	minimal	low	medium	high	Source
GPQA Diamond	85.9%#84	86.4%#78	–	94.1%#10	Epoch AI Benchmarking Hub
Arena Hard Prompts	–	–	–	1501#31	Arena (LMArena)
LiveBench 1.0 Reasoningnot in index	–	–	–	85.2%#38	LiveBench
GPQA Diamond (AA)not in index	–	–	–	92.8%#28	Artificial Analysis
Humanity's Last Exam (AA)not in index	–	–	–	40.8%#83	Artificial Analysis
GPQA Diamond (Vals)not in index	–	–	–	93.4%#7	Vals AI
ARC-AGI-1	34.5%#199	76.5%#129	83.2%#117	91.2%#67	ARC Prize
ARC-AGI-2	2.6%#197	30.4%#132	50.4%#108	60.4%#86	ARC Prize
CritPt	–	–	–	10.6%#117	Artificial Analysis
Chess Puzzlesnot in index	29.0%#69	42.0%#36	–	40.0%#42	Epoch AI Benchmarking Hub
Mystery Game Puzzlesnot in index	25.0%#47	23.0%#51	–	30.0%#34	Epoch AI Benchmarking Hub
LMCAnot in index	–	–	–	44.9%#91	Epoch AI Benchmarking Hub
DTBenchnot in index	–	–	–	95.5%#38	Epoch AI Benchmarking Hub
ForecastBenchnot in index	–	–	65.9%#8	–	ForecastBench

Coding

Benchmark	high	not stated	Source
SciCode	53.4%#78	–	SciCode
WeirdML	56.1%#57	–	WeirdML
FrontierCodeeffort not stated	–	34.4%#30	Cognition
Arena Coding	1518#41	–	Arena (LMArena)
Arena WebDev	1538#46	–	Arena (LMArena)
LiveBench 1.0 Codingnot in index	77.9%#35	–	LiveBench
SciCode (AA)not in index	53.4%#77	–	Artificial Analysis
LiveCodeBench	88.1%#8	–	Vals AI
IOI	35.1%#38	–	Vals AI
SWE-bench (Vals)not in index	79.6%#25	–	Vals AI
Code Migrationnot in index	30.9%#40	–	Vals AI
ProgramBenchnot in index	0.0%#32	–	Vals AI
Vibe Code Bench v1.1not in index	64.0%#50	–	Vals AI
DeepSWE v1.1not in index	46.7%#51	–	Epoch AI Benchmarking Hub
ALE-Benchnot in index	715.5#85	–	Epoch AI Benchmarking Hub
Terminal-Bench 4.0 (AA)not in index	7.1%#95	–	Artificial Analysis
Terminal-Bench 2.1 (AA)not in index	77.5%#57	–	Artificial Analysis

Agents & tools

Benchmark	high	not stated	Source
APEX-Agentseffort not stated	–	46.9%#34	Mercor
LiveBench Agentic Codingnot in index	43.4%#57	–	LiveBench
Terminal-Bench 2.1 (Vals)	73.8%#21	–	Vals AI
τ³-Banking (AA)	29.9%#67	–	Artificial Analysis
CyberBenchnot in index	45.4%#39	–	Vals AI
Finance Agent v2not in index	56.3%#21	–	Vals AI
Harvey's Legal Agent Benchmark v1not in index	3.3%#42	–	Vals AI
Legal Research Benchnot in index	25.0%#57	–	Vals AI
Tax Agent Benchnot in index	18.0%#50	–	Vals AI
Terminal-Bench Sciencenot in index	4.3%#25	–	Vals AI
AutomationBenchnot in index	53.0%#70	–	Artificial Analysis
GDP.pdf (AA)not in index	17.4%#81	–	Artificial Analysis
FACTS Searchnot in indexeffort not stated	–	68.2%#18	Google

Maths

Benchmark	minimal	low	high	not stated	Source
FrontierMath Tiers 1–3	–	–	59.0%#42	–	Epoch AI Benchmarking Hub
FrontierMath Tier 4	–	–	21.9%#52	–	Epoch AI Benchmarking Hub
OTIS Mock AIME	80.0%#114	82.2%#108	94.2%#56	–	Epoch AI Benchmarking Hub
ProofBencheffort not stated	–	–	–	36.0%#39	Vals AI
LiveBench Mathematicsnot in index	–	–	86.4%#51	–	LiveBench
AIME 2026effort not statedpossible contamination	–	–	–	96.7%#6	MathArena
HMMT February 2026effort not statedpossible contamination	–	–	–	89.4%#14	MathArena
MathArena Apexeffort not statedpossible contamination	–	–	–	26.0%#12	MathArena
Arena Maths	–	–	1504#13	–	Arena (LMArena)

Knowledge

Benchmark	high	not stated	Source
SimpleQA Verified	66.2%#12	–	Epoch AI Benchmarking Hub
LiveBench 1.0 Data Analysisnot in index	63.0%#59	–	LiveBench
AA-Omniscience	22.1#58	–	Artificial Analysis
MMLU-Pro	89.3%#13	–	Vals AI
LegalBench	86.7%#12	–	Vals AI
CorpFin	63.3%#45	–	Vals AI
TaxEval	74.9%#25	–	Vals AI
BioMysteryBenchnot in index	58.5%#26	–	Vals AI
Excel Modeling Benchmarknot in index	65.4%#25	–	Vals AI
MedCodenot in index	53.1%#11	–	Vals AI
MedScribenot in index	79.7%#60	–	Vals AI
Public Benefits Bench v1.1not in index	56.6%#39	–	Vals AI
GDP.pdfnot in index	14.0%#42	14.0%#42	Epoch AI Benchmarking Hub
AA-Omniscience: accuracynot in index	50.0%#62	–	Artificial Analysis
AA-Omniscience: non-hallucinationnot in index	44.4%#145	–	Artificial Analysis

Instruction following

Benchmark	high	Source
LiveBench Languagenot in index	83.9%#17	LiveBench
LiveBench Instruction Followingnot in index	75.4%#9	LiveBench
Arena Instruction Followingnot in index	1473#32	Arena (LMArena)

Human preference

Benchmark	high	Source
Arena Text	1483#22	Arena (LMArena)
Arena Creative Writingnot in index	1471#15	Arena (LMArena)
Arena Multi-turnnot in index	1488#24	Arena (LMArena)
Arena Longer Queriesnot in index	1485#37	Arena (LMArena)

Multimodal

Benchmark	high	not stated	Source
Arena Vision	1299#20	–	Arena (LMArena)
MMMU-Pro	83.2%#27	–	Artificial Analysis
MMMU-Pro (Vals)not in index	88.4%#7	–	Vals AI
MortgageTaxnot in index	67.6%#24	–	Vals AI
SAGEnot in index	48.7%#25	–	Vals AI
Blueprint-Bench 2not in indexeffort not stated	–	31.2%#15	Epoch AI Benchmarking Hub

Long context

Benchmark	high	Source
AA-LCR	80.0%#78	Artificial Analysis
Arena Documentnot in index	1453#28	Arena (LMArena)
MLCRnot in index	10.0%#113	Artificial Analysis

Composite

Benchmark	high	not stated	Source
Epoch Capabilities Indexnot in indexeffort not stated	–	154.5#40	Epoch AI Benchmarking Hub
LiveBench 1.0not in index	73.6%#44	–	LiveBench
AA Intelligence Index v4.3.2	34.0#99	–	Artificial Analysis
Vals Multimodal Indexnot in index	65.1%#13	–	Vals AI

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	195 tok/s	1.58 s	$1.35	$6.75	1.0M	–
Google Vertex	153 tok/s	2.15 s	$1.35	$6.75	1.0M	–
Google AI Studio	140 tok/s	1.65 s	$0.750	$3.75	1.0M	–
Google AI Studio	112 tok/s	1.14 s	$0.375	$1.88	1.0M	–
Google Vertex	108 tok/s	1.35 s	$0.750	$3.75	1.0M	–
Google Vertex	105 tok/s	0.81 s	$0.825	$4.13	1.0M	–
Google Vertex	76 tok/s	7.99 s	$0.375	$1.88	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 $0.075 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0014	$0.0012	17.2 s
Summarize a 30-page report	12,000 / 600	$0.011	$0.0052	18.7 s
Code edit	6,000 / 1,500	$0.010	$0.0071	23.2 s
Agentic coding session	60,000 / 4,000	$0.060	$0.030	35.5 s
Structured extraction	2,000 / 200	$0.0023	$0.0012	16.7 s

New this month

Compare Gemini 3.6 Flash with…

Questions people ask

What are Gemini 3.6 Flash's benchmark scores?: As of 24 Oct 2026, Gemini 3.6 Flash (high effort) scores 62.0 ±2.9 on the BenchLeader Index, top 100, strongest in composite (68) and weakest in coding (57). That rests on 28 independent benchmark results, and every score on this page links to its publisher.
How does Gemini 3.6 Flash rank?: Gemini 3.6 Flash (high effort) is #89 of 785 ranked configurations on 24 Oct 2026 (top 100). 170 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 3.6 Flash cost per 1M tokens?: As of 24 Oct 2026, Gemini 3.6 Flash costs $0.750 per 1M input tokens and $3.75 per 1M output tokens, $1.50 blended (three input tokens per output token). That is a launch price until 31 Dec 2026, then $3.00 blended.
Is Gemini 3.6 Flash better than GPT-5.6 Sol?: As of 24 Oct 2026, GPT-5.6 Sol (max) leads on quality (68.0 vs 62.0 on the BenchLeader Index), but Gemini 3.6 Flash (high) costs 5.3× less, so it is the better buy unless you need the extra quality. The head-to-head page has the benchmark-by-benchmark detail.
When was Gemini 3.6 Flash released?: Google released Gemini 3.6 Flash 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 3.6 Flash: benchmarks, pricing, speed and rank”, data as of 24 Oct 2026, https://www.benchleader.com/models/gemini-3-6-flash.