VerdictPal
Join

Models / Google

Gemini 3.5 Flash

Gemini · Released 2026-05-19

Google's fastest frontier-tier model. Leads the MCP Atlas tool-use benchmark (83.6%) and pairs a 1M-token context with sub-200ms first-token latency. Terminal-Bench 2.0 76.2%, ARC-AGI-2 72.1%. $0.30 / $1.20 per 1M tokens.

Default model for high-throughput, low-latency tool use: search agents, batch classification, real-time customer-facing apps. The cost-per-1M gap to Opus 4.8 is roughly 50× on input and 60× on output.

#16 of 24 on Vals Index · current snapshot
49 Vals Index · 30 benchmark rows · 11 sourcesPosition is benchmark-specific — not a cross-family or cross-source ranking.
Composite66
Context1M
Input / 1M$0.30
Output / 1M$1.2
Knowledge cutoff2026-03-01
Statussolid
GeminiNotebookLMScira AI

Family profile

Best published score in each covered benchmark family.

79/100 avg
Knowledge
88
Reasoning
76
Math
39
Coding
95
Agentic
76
Multimodal
84
Long context
81
Tool use
84
Human preference
91

9 tested benchmark families

Editor's note
3.5 Flash is the model we recommend for high-volume tool use. The cost-per-1M gap to Opus 4.8 is roughly 50× on input and 60× on output — a meaningful line item for any team running agents at scale. The MCP Atlas 83.6% makes it the strongest tool-use model in the public market.
Benchmark placements

Where Gemini 3.5 Flash places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.

BenchmarkFamilyRankScoreSourceDate
MCP AtlasTool use#1/ 1084MCP Atlas leaderboard2026-05-15
LMArena (Chatbot Arena)Human preference#3/ 1391LMArena leaderboard2026-05-25
MMMU-ProMultimodal#3/ 884Artificial Analysis — MMMU-Pro2026-09-01
AA-LCRLong context#4/ 17981Artificial Analysis2026-09-01
MMMUMultimodal#4/ 869MMMU leaderboard2026-05-19
AA-OmniscienceKnowledge#6/ 1252Artificial Analysis — AA-Omniscience2026-09-01
BFCLTool use#6/ 2082Berkeley Function Calling Leaderboard V42026-04-15
HumanEvalCoding#6/ 1595CodeSOTA HumanEval leaderboard2026-05-19
ARC-AGIReasoning#7/ 1472ARC Prize leaderboard2026-05-15
LongBench v2Long context#7/ 1954LongBench v2 leaderboard2026-04-01
DeepSWECoding#8/ 837DeepSWE leaderboard v1.12026-06-20
LiveBenchReasoning#8/ 2164LiveBench leaderboard2026-04-15
FrontierMathMath#10/ 1539Epoch AI FrontierMath Tiers 1-3 (v2) CSV2026-08-15
MMLUKnowledge#12/ 2288Google Gemini 3.5 Flash2026-05-19
IFBenchReasoning#17/ 12776Artificial Analysis2026-09-01
Humanity's Last ExamReasoning#20/ 18043Artificial Analysis2026-09-01
SWE-bench VerifiedCoding#20/ 2377SWE-bench Verified leaderboard2026-06-09
SciCodeCoding#26/ 18053Artificial Analysis2026-09-01
Artificial Analysis Intelligence IndexReasoning#29/ 18752Artificial Analysis2026-09-01
τ³-Bench BankingAgentic#33/ 10632Artificial Analysis2026-09-01
Terminal-BenchAgentic#34/ 18476Terminal-Bench 2.0 leaderboard2026-05-20
AA time to first tokenPerformance#116/ 186Artificial Analysis2026-09-01
AA output speedPerformance#117/ 187Artificial Analysis2026-09-01
GPQA DiamondReasoning#158/ 18574Artificial Analysis GPQA Diamond evaluation2026-06-09
Family context

The three highest-scoring models with pages in each capability family. Where Gemini 3.5 Flash shows up, it's highlighted.

KnowledgeGemini 3.5 Flash · 88
  1. 1
    GPT 5.5
    91
  2. 2
    Claude Opus 4.8
    91
  3. 3
    Claude Sonnet 4.6
    91
ReasoningGemini 3.5 Flash · 76
  1. 1
    Claude Mythos Preview
    95
  2. 2
    Gemini 3.7 Flash
    95
  3. 3
    Grok 4.6
    95
MathGemini 3.5 Flash · 39
  1. 1
    OpenAI o3
    99
  2. 2
    GPT-5 (medium)
    99
  3. 3
    GPT-5 (high)
    99
CodingGemini 3.5 Flash · 95
  1. 1
    GPT 5.5
    96
  2. 2
    Claude Opus 4.8
    96
  3. 3
    Claude Opus 4.6
    96
AgenticGemini 3.5 Flash · 76
  1. 1
    GPT 5.5
    100
  2. 2
    Claude Opus 4.8
    100
  3. 3
    Claude Fable 5
    100
Long contextGemini 3.5 Flash · 81
  1. 1
    Muse Spark 1.2
    83
  2. 2
    Kimi K3
    83
  3. 3
    Gemini 3.5 Flash
    81
Tool useGemini 3.5 Flash · 84
  1. 1
    Claude Sonnet 4.6
    89
  2. 2
    GPT 5.5
    87
  3. 3
    Claude Opus 4.8
    87
Human preferenceGemini 3.5 Flash · 91
  1. 1
    GPT 5.5
    95
  2. 2
    Claude Opus 4.8
    92
  3. 3
    Gemini 3.5 Flash
    91
Newest receipts

Benchmark rows added to the public ledger for Gemini 3.5 Flash in the last 120 days. Older rows live in the full table below.

23 recent
BenchmarkFamilyScoreSourceDays ago
AA-OmniscienceKnowledge51.9%Artificial Analysis — AA-Omniscience6
MMMU-ProMultimodal84.3%Artificial Analysis — MMMU-Pro6
AA-LCRLong context81%Artificial Analysis6
AA output speedPerformance0 t/sArtificial Analysis6
AA time to first tokenPerformance0.00sArtificial Analysis6
Artificial Analysis Intelligence IndexReasoning52%Artificial Analysis6
Humanity's Last ExamReasoning42.7%Artificial Analysis6
IFBenchReasoning76.33%Artificial Analysis6
SciCodeCoding53.1%Artificial Analysis6
τ³-Bench BankingAgentic32.16%Artificial Analysis6
FrontierMathMath38.97%Epoch AI FrontierMath Tiers 1-3 (v2) CSV23
DeepSWECoding37%DeepSWE leaderboard v1.179
SWE-bench VerifiedCoding77.2%SWE-bench Verified leaderboard90
GPQA DiamondReasoning74.1%Artificial Analysis GPQA Diamond evaluation90
Vals IndexAgentic49.31%Vals AI — Vals Index95
SWE-bench VerifiedCoding78%SWE-bench official leaderboard100
LMArena (Chatbot Arena)Human preference1455LMArena leaderboard105
Terminal-BenchAgentic76.2%Terminal-Bench 2.0 leaderboard110
HumanEvalCoding94.6% pass@1CodeSOTA HumanEval leaderboard111
MMLUKnowledge88.4%Google Gemini 3.5 Flash111
MMMUMultimodal68.9%MMMU leaderboard111
ARC-AGIReasoning72.1%ARC Prize leaderboard115
MCP AtlasTool use83.6%MCP Atlas leaderboard115
Family coverage

A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.

Knowledge88
Reasoning76
Math39
Coding95
Agentic76
Multimodal84
Long context81
Tool use84
Human preference91
Price vs. performance
0255075100$0.10$0.30$1$3$10$30input USD / 1M tokens · log scale →← intelligence

Hover any dot for name, score, and input price. Keyboard: tab through the top twelve, or use the ranking below.

AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMetaMiniMaxxAIZhipuXiaomiOtherMeituanCohereNVIDIA

Full benchmark ledger

Every catalog benchmark for Gemini 3.5 Flash. Scores link to the original source; gaps mean no public row exists yet.

30 sourced rows

24 of 37 catalog benchmarks have a sourced row for Gemini 3.5 Flash.

Knowledge

Reasoning

Math

Coding

Agentic

Multimodal

Long context

Tool use

Performance

Human preference

Changelog
  1. Released Gemini 3.5 Flash — MCP Atlas leader (83.6%), 1M context, cheapest frontier pricing.
Sources