GoogleReasoning modelCompare with another model

Gemini 3.5 Flash

Reasoning effort

Gemini 3.5 Flash is a Google proprietary reasoning model, released 19 May 2026. Its best configuration (high reasoning effort) ranks #68 of 775 on the BenchLeader Index at 63.2 ±2.8, in the upper half. The ± is the point: 143 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 instruction following (74) and lowest in maths (58). At $3.38 per million tokens blended it is pricier than most ranked models. Output speed of 208 tokens per second puts it in the fastest quarter, with a first answer in 15.5 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 19 Oct 2026.

Blended price: $3.38/M; $1.50 in · $9.00 out
Output speed: 208 tok/s; measured by Artificial Analysis
First answer: 15 s; first token 1.64 s
Context: 1.0M
Full answer: 18 s; median, reasoning included
Cost per run: $3.69; one full Intelligence Index run
Released: 19 May 2026

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

Overall index63
Reasoning61
Coding60
Agents & tools63
Maths58
Knowledge65
Instruction following74
Human preference67
Multimodal67
Long context62
Composite66

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	64.5	#48
Gemini 3.7 Flashhigh	13 Aug 2026	63.5	#61
Gemini 3.6 Flashhigh	21 Jul 2026	62.1	#89
Gemini 3.5 Flashhighthis page	19 May 2026	63.2	#68
Gemini 3 Flashthinking	17 Dec 2025	59.2	#134
Gemini 2.5 Flash 09 2025thinking	25 Sept 2025	51.1	#348
Gemini 2.5 Flash	17 Apr 2025	52.7	#302
Gemini 2.0 Flash	5 Feb 2025	43.3	#560
Gemini 1.5 Flash 002	24 Sept 2024	40.2	#650
Gemini 1.5 Flash 001	23 May 2024	37.2	#726
Gemini 1.5 Flash	14 May 2024	–	–
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	56.4	#203	188 tok/s	0.84 s	$0.0033	Agents & tools 74 · Composite 56 · Instruction following 49 · Knowledge 62 · Long context 55 · Maths 59 · Multimodal 63 · Reasoning 47
low	–	–	114 tok/s	1.64 s	$0.0033	Maths 62 · Reasoning 64
medium	62.9	#77	199 tok/s	12 s	$0.0033	Agents & tools 68 · Coding 57 · Composite 67 · Human preference 67 · Instruction following 73 · Knowledge 71 · Long context 62 · Maths 67 · Multimodal 66 · Reasoning 61
highbest	63.2	#68	208 tok/s	15 s	$0.0033	Agents & tools 63 · Coding 60 · Composite 66 · Human preference 67 · Instruction following 74 · Knowledge 65 · Long context 62 · Maths 58 · Multimodal 67 · Reasoning 61
not stated	49.1	#402	114 tok/s	1.64 s	$0.0033	Agents & tools 32 · Human preference 33 · Maths 58 · Reasoning 72

How its index has moved

8 Sept 2026 to 21 Oct 2026

high
medium
minimal
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	not stated	Source
GPQA Diamond	86.4%#77	88.9%#54	–	92.8%#23	–	Epoch AI Benchmarking Hub
SimpleBencheffort not stated	–	–	–	–	76.7%#6	SimpleBench
Arena Hard Prompts	–	–	1495#44	1495#41	–	Arena (LMArena)
LiveBench 1.0 Reasoningnot in index	–	–	–	82.0%#48	–	LiveBench
GPQA Diamond (AA)not in index	82.8%#163	–	92.1%#39	92.2%#38	–	Artificial Analysis
Humanity's Last Exam (AA)not in index	24.1%#186	–	41.3%#76	42.7%#63	–	Artificial Analysis
GPQA Diamond (Vals)not in index	–	–	–	92.7%#14	–	Vals AI
EnigmaEval	–	–	–	25.4%#4	–	Scale AI SEAL
ARC-AGI-1	48.8%#168	–	–	92.5%#51	–	ARC Prize
ARC-AGI-2	8.9%#152	–	–	72.1%#52	–	ARC Prize
CritPt	1.4%#207	–	10.9%#113	13.1%#102	–	Artificial Analysis
Chess Puzzlesnot in index	27.0%#71	42.0%#35	–	45.0%#28	–	Epoch AI Benchmarking Hub
Mystery Game Puzzlesnot in index	–	28.0%#38	–	32.0%#29	–	Epoch AI Benchmarking Hub
LMCAnot in index	–	–	–	47.1%#80	–	Epoch AI Benchmarking Hub
DTBenchnot in index	–	–	–	94.7%#52	–	Epoch AI Benchmarking Hub
ForecastBenchnot in indexeffort not stated	–	–	–	–	64.3%#28	ForecastBench

Coding

Benchmark	medium	high	Source
SWE-bench Verified (Epoch)	–	79.3%#3	Epoch AI Benchmarking Hub
SciCode	–	53.9%#69	SciCode
WeirdML	–	62.6%#44	WeirdML
Arena Coding	1506#66	1508#61	Arena (LMArena)
Arena WebDev	1491#60	1499#58	Arena (LMArena)
LiveBench 1.0 Codingnot in index	–	78.2%#32	LiveBench
SciCode (AA)not in index	–	53.9%#72	Artificial Analysis
LiveCodeBench	–	87.6%#12	Vals AI
SWE-bench (Vals)not in index	–	78.8%#28	Vals AI
Code Migrationnot in index	–	26.8%#45	Vals AI
ProgramBenchnot in index	–	0.0%#32	Vals AI
Vibe Code Bench v1.1not in index	–	48.7%#59	Vals AI
Surface Evolver Benchnot in index	58.1%#14	–	Epoch AI Benchmarking Hub
DeepSWE v1.1not in index	37.4%#58	36.1%#60	Epoch AI Benchmarking Hub
ALE-Benchnot in index	–	911.0#64	Epoch AI Benchmarking Hub
Terminal-Bench 4.0 (AA)not in index	–	6.6%#96	Artificial Analysis
Terminal-Bench 2.1 (AA)not in index	–	78.7%#49	Artificial Analysis

Agents & tools

Benchmark	minimal	medium	high	not stated	Source
APEX-Agentseffort not stated	–	–	–	27.5%#50	Mercor
LiveBench Agentic Codingnot in index	–	–	49.0%#46	–	LiveBench
Terminal-Bench Hard	46.2%#30	39.4%#57	40.9%#53	–	Artificial Analysis
τ²-Bench Telecom (AA)not in index	58.8%#182	95.6%#19	95.3%#23	–	Artificial Analysis
Terminal-Bench 2.1 (Vals)	–	–	74.2%#20	–	Vals AI
MCP Atlas	–	–	83.6%#6	–	Scale AI SEAL
HiL-Bencheffort not stated	–	–	–	27.7%#24	Scale AI SEAL
GDPval-AA v2.1not in index	–	–	35.1%#119	–	Artificial Analysis
τ³-Banking (AA)	–	–	32.2%#58	–	Artificial Analysis
ITBench SRE (AA)not in index	–	–	40.4%#17	–	Artificial Analysis
Analyst Agent (AA)not in index	–	–	45.0%#17	–	Artificial Analysis
APEX-Agents (AA)not in index	–	–	47.0%#1	–	Artificial Analysis
Finance Agent v2not in index	–	–	57.9%#13	–	Vals AI
Legal Research Benchnot in index	–	–	30.8%#46	–	Vals AI
SkillsBenchnot in index	–	–	52.7%#18	–	Vals AI
Tax Agent Benchnot in index	–	–	21.6%#43	–	Vals AI
Terminal-Bench 2.0 (Vals)not in index	–	–	67.4%#4	–	Vals AI
Terminal-Bench 4.0 (Vals)not in index	–	–	6.1%#41	–	Vals AI
Terminal-Bench Sciencenot in index	–	–	5.7%#21	–	Vals AI
Web Search Indexnot in index	–	–	37.0%#7	–	Vals AI
GBAEvalnot in indexeffort not stated	–	–	–	6.7%#14	Epoch AI Benchmarking Hub
Vending-Bench 2not in indexeffort not stated	–	–	–	5396.4#28	Andon Labs
AutomationBenchnot in index	–	–	42.1%#92	–	Artificial Analysis
GDP.pdf (AA)not in index	–	–	19.8%#66	–	Artificial Analysis
EnterpriseOps-Gymnot in index	–	–	50.1%#3	–	Artificial Analysis
AA-Briefcase v1.1not in index	–	–	874#53	–	Artificial Analysis
FACTS Searchnot in indexeffort not stated	–	–	–	64.3%#19	Google

Maths

Benchmark	minimal	low	medium	high	not stated	Source
FrontierMath Tiers 1–3	–	–	–	62.8%#38	–	Epoch AI Benchmarking Hub
FrontierMath Tier 4	–	–	–	26.8%#42	–	Epoch AI Benchmarking Hub
OTIS Mock AIME	80.0%#113	88.9%#75	–	95.6%#43	–	Epoch AI Benchmarking Hub
ProofBench	–	–	–	31.0%#42	–	Vals AI
LiveBench Mathematicsnot in index	–	–	–	88.2%#43	–	LiveBench
AIME 2026effort not statedpossible contamination	–	–	–	–	95.0%#18	MathArena
HMMT February 2026effort not statedpossible contamination	–	–	–	–	95.5%#6	MathArena
MathArena Apexeffort not statedpossible contamination	–	–	–	–	32.3%#9	MathArena
Arena Maths	–	–	1482#35	1497#19	–	Arena (LMArena)

Knowledge

Benchmark	minimal	medium	high	not stated	Source
SimpleQA Verified	–	–	66.2%#12	–	Epoch AI Benchmarking Hub
LiveBench 1.0 Data Analysisnot in index	–	–	64.9%#58	–	LiveBench
AA-Omniscience	0.7#130	20.8#64	21.2#61	–	Artificial Analysis
MMLU-Pro	–	–	89.5%#10	–	Vals AI
LegalBench	–	–	83.6%#51	–	Vals AI
CorpFin	–	–	64.7%#35	–	Vals AI
TaxEval	–	–	74.4%#36	–	Vals AI
Excel Modeling Benchmarknot in index	–	–	63.5%#27	–	Vals AI
MedCodenot in index	–	–	55.8%#6	–	Vals AI
MedScribenot in index	–	–	76.6%#73	–	Vals AI
Public Benefits Benchnot in index	–	–	59.5%#32	–	Vals AI
DrugDiscoveryBenchnot in index	–	–	50.0%#7	–	Scale AI SEAL
EBR-benchnot in index	–	–	4.8%#25	–	Epoch AI Benchmarking Hub
GDP.pdfnot in index	–	14.0%#42	–	14.0%#42	Epoch AI Benchmarking Hub
AA-Omniscience: accuracynot in index	43.0%#101	51.0%#59	51.4%#57	–	Artificial Analysis
AA-Omniscience: non-hallucinationnot in index	25.7%#237	38.2%#170	37.8%#171	–	Artificial Analysis

Instruction following

Benchmark	minimal	medium	high	Source
LiveBench Languagenot in index	–	–	84.6%#16	LiveBench
LiveBench Instruction Followingnot in index	–	–	75.6%#7	LiveBench
IFBench	47.3%#187	74.6%#35	76.3%#21	Artificial Analysis
Arena Instruction Followingnot in index	–	1466#42	1465#43	Arena (LMArena)

Human preference

Benchmark	medium	high	not stated	Source
Arena Text	1476#33	1478#29	–	Arena (LMArena)
EQ-Bench 4effort not stated	–	–	1087#23	EQ-Bench
Arena Creative Writingnot in index	1468#21	1464#24	–	Arena (LMArena)
Arena Multi-turnnot in index	1478#45	1480#40	–	Arena (LMArena)
Arena Longer Queriesnot in index	1480#48	1482#39	–	Arena (LMArena)

Multimodal

Benchmark	minimal	medium	high	not stated	Source
Arena Vision	–	1306#14	1308#13	–	Arena (LMArena)
MMMU-Pro	80.1%#52	83.9%#24	84.3%#20	–	Artificial Analysis
MMMU-Pro (Vals)not in index	–	–	88.3%#8	–	Vals AI
MortgageTaxnot in index	–	–	68.1%#19	–	Vals AI
SAGEnot in index	–	–	49.9%#18	–	Vals AI
Blueprint-Bench 2not in indexeffort not stated	–	–	–	33.6%#10	Epoch AI Benchmarking Hub

Long context

Benchmark	minimal	medium	high	Source
AA-LCR	61.3%#270	74.3%#160	73.3%#175	Artificial Analysis
Arena Documentnot in index	–	1461#25	1453#28	Arena (LMArena)
MLCRnot in index	–	–	18.3%#33	Artificial Analysis

Composite

Benchmark	minimal	medium	high	not stated	Source
Epoch Capabilities Indexnot in indexeffort not stated	–	–	–	154.9#38	Epoch AI Benchmarking Hub
LiveBench 1.0not in index	–	–	74.6%#39	–	LiveBench
AA Intelligence Index v4.3.2	23.9#179	33.6#104	32.6#109	–	Artificial Analysis
Vals Multimodal Indexnot in index	–	–	62.9%#16	–	Vals AI

Where it wins

Benchmarks where this configuration ranks in the top five of every configuration measured.

#3SWE-bench Verified (Epoch)79.3% · of 33

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	164 tok/s	1.79 s	$1.50	$9.00	1.0M	–
Google Vertex	136 tok/s	1.56 s	$1.50	$9.00	1.0M	–
Google AI Studio	111 tok/s	0.89 s	$0.750	$4.50	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 $0.150 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0033	$0.0029	16.9 s
Summarize a 30-page report	12,000 / 600	$0.023	$0.011	18.4 s
Code edit	6,000 / 1,500	$0.022	$0.016	22.7 s
Agentic coding session	60,000 / 4,000	$0.126	$0.065	34.7 s
Structured extraction	2,000 / 200	$0.0048	$0.0028	16.4 s