Model

Gemini 3.5 Flash

CLOSED

Google · Gemini 3.5 family · released May 19, 2026

First Flash to beat the prior-gen Pro flagship; I/O 2026 headline, 4x faster output than rival frontier models.

ReasoningCodingVisionFunction callingTool useAgentic

2481.7

Elo · rank #43

Parameters

Undisclosed

Active params

Undisclosed

Context

1M tokens

Architecture

Sparse Mixture-of-Experts Transformer (configuration undisclosed)

License

Proprietary

Languages

—

API price (in/out)

$1.5 / $9

Modalities

text · vision · audio · video

Benchmark results

Bar shows position within the tracked field; marker = field best

AIMEMath95.0%#17

best: GPT-5.2 · 100.0%

ARC-AGI-2Reasoning72.1%#14

best: GPT-6 Astra · 95.0%

Arena EloHuman preference1478#18

best: Claude Fable 5 · 1504

CharXivVision84.2%#8

best: Qwen3.8-Flash-Next · 90.6%

GDPval-AAAgents1656#10

best: Claude Fable 5 · 1932

GPQA DiamondReasoning92.2%#21

best: GPT-6 Astra · 96.0%

Humanity's Last ExamReasoning40.2%#33

best: Claude Opus 5.5 · 67.7%

IFBenchReasoning76.3%#16

best: MiniMax M3 · 83.0%

LiveCodeBenchCoding87.6%#17

best: DeepSeek-V4-Pro (Think Max) · 93.5%

MMMU-ProVision83.6%#2

best: Claude Opus 4.7 · 85.5%

MMMUVision83.0%#14

best: Claude Fable 5 · 89.3%

OSWorld-VerifiedAgents78.4%#10

best: Claude Fable 5 · 85.0%

SWE-bench ProCoding55.1%#37

best: Claude Opus 5.5 · 89.9%

SWE-bench VerifiedCoding78.8%#33

best: Claude Opus 5 · 97.0%

Terminal-Bench 2.0Coding76.2%#25

best: DeepSeek-V4.1-Flash · 90.6%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$1.5

Output / M tok

Gemini 3.5 family

Elo progression across releases

Gemini 3.5 FlashMay 20262481.7 Gemini 3.5 Flash CyberJul 20262712.6 Gemini 3.5 Flash-LiteJul 20262073.9

Resources

DOCShttps://deepmind.google/models/gemini/flash/APIAvailable via Google API

API price $1.5/$9 · each benchmark row carries its own source badge (see methodology)