Google logo

Gemini 3.5 Flash

Google · Released May 19, 2026

ProprietarycodingagentsbudgetCompare this model

$1.5

Input price / MTok

#16 of 27

$9

Output price / MTok

#18 of 27

1.0M

Context window

#10 of 28

198

Tokens / sec

#3 of 21

55.1

SWE-bench Pro

#9 of 9

78.4

OSWorld

#5 of 8

77.3

MRCR

#3 of 5

Summary

Gemini 3.5 Flash is a proprietary language model from Google, released on May 19, 2026.

At $1.5 input and $9 output per million tokens, it ranks #16 of 27 for input price, placing it mid-range on input cost.

Its context window of 1.0M tokens ranks #10 of 28 for context depth in this dataset.

It targets coding, agents, budget workloads.

Throughput is measured at 198 tokens per second (ranked #3 of 21 for throughput).

Launched at Google I/O 2026; outperforms Gemini 3.1 Pro on coding benchmarks.

Benchmarks

BenchmarkScoreMax%RankSource
SWE-bench Pro55.110055.1%#9/9Google DeepMind model card (self-reporte...
OSWorld78.410078.4%#5/8Google DeepMind model card (self-reporte...
MRCR77.310077.3%#3/5Google DeepMind model card (self-reporte...

Similar Models

Compare all 6 models

Models with the most shared benchmark coverage and closest scores to Gemini 3.5 Flash.

Gemini 3.5 Flash
OpenAIGPT-5.5
AnthropicClaude Sonnet 4.6
OpenAIGPT-5.4
AnthropicClaude Sonnet 5
AnthropicClaude Opus 4.8
SWE-bench Pro#9 of 9
Professional software engineering tasks
OSWorld#5 of 8
Computer-use and GUI automation
MRCR#3 of 5
Long-context retrieval and recall

Compare Gemini 3.5 Flash side by side

Pre-filled with 5 similar models

Open comparison

Speed source: https://artificialanalysis.ai/models/gemini-3-5-flash