Glimmer Index
Plus $4/mo

Gemini 3.5 Flash

Google · Released May 19, 2026 · Closed weights

Fast Gemini model balancing multimodal reasoning, tool use, and cost

Strongest at
Visual reasoning
Top 22% of the past year's models · 2 of 3 benchmarks
Price per 1M tokens
$1.5 / $9
input / output · Google
Context window
1.05M
tokens in · 65.5K out
Since the last one
153 days
after Gemini 3 Flash Preview
Lab announcement ↗Open in the timeline

Capability profile

Scored against 91 models released in the past year. 50 is the average model; further out is better.

This modelGemini 3 Flash PreviewMedian of the past year
5090Median of the past yearCoding: 47 · Better than 25%Writing: 56 · Better than 47%Reading & documents: 56 · Top 41%Reasoning: 64 · Top 35%Math: 54 · Better than 41%Science: 65 · Top 42%Visual reasoning: 67 · Top 22%Computer use: 53 · Better than 10%Gemini 3 Flash Preview: 49Gemini 3 Flash Preview: 49Gemini 3 Flash Preview: 65Gemini 3 Flash Preview: 58Gemini 3 Flash Preview: 50Gemini 3 Flash Preview: 63Gemini 3 Flash Preview: 54CodingWritingReadingReasoningMathScienceVisionComputer use
Coding47−2.2Better than 25%
Writing56+6.8Better than 47%
Reading & documents56−9.3Top 41%
Reasoning64+6.0Top 35%
Math54+4.0Better than 41%
Science65+1.5Top 42%
Visual reasoning67+13.3Top 22%
Computer use53Better than 10%
Where these scores come from

Each capability combines the published benchmarks listed here, rescaled so 50 is the reference average and 15 points is one standard deviation. A hollow dot rests on a single benchmark.

Price against capability

The shaded corner holds anything better and cheaper. Prices blend input and output 3 to 1.

2 models from the past year score higher on Visual reasoning for less: GPT-6.1 Terra, Gemini 3.7 Flash.

$0.1$0.3$1$3$1020406080GPT-6.1 Terra: 79, $2 per 1MMuse Spark Open: 58, $2 per 1MClaude Haiku 5.5: 59, $0.2 per 1MMistral Large 4: 42, $2.06 per 1MGemini 4 Argon: 76, $4 per 1MGPT-6.1 Sol: 79, $4 per 1MClaude Sonnet 5.5: 71, $4 per 1MGPT-6 Sol: 66, $4 per 1MGPT-6 Luna: 59, $0.2 per 1MClaude Opus 5.5: 76, $8 per 1MGrok 4.7: 56, $3 per 1MDeepSeek V4.1 Flash: 54, $0.21 per 1MGPT-6 Astra: 78, $20 per 1MMuse Spark 1.3: 66, $2 per 1MGemini 3.8 Flash: 66, $1.5 per 1MClaude Fable 5.1: 68, $20 per 1MQwen3.8 Flash: 46, $0.23 per 1MGLM-5.3-Flash: 41, $0.24 per 1MDeepSeek V4 Flash Vision Exp: 22, $0.32 per 1MQwen3.8 27B: 29, $1.12 per 1MGemini 3.7 Flash: 67, $1.5 per 1MGrok 4.6: 51, $3 per 1MMuse Glimmer 30B: 43, $0.52 per 1MMuse Spark 1.2: 56, $2 per 1MQwen3.8 Max: 66, $3 per 1MClaude Opus 5: 60, $10 per 1MGemini 3.6 Flash: 57, $1.5 per 1MGemini 3.5 Flash Lite: 41, $0.85 per 1MKimi K3: 49, $6 per 1MQwen3.7 Flash: 30, $0.055 per 1MGPT-5.6 Terra: 55, $4.5 per 1MGPT-5.6 Sol: 66, $8 per 1MGPT-5.6 Luna: 51, $0.45 per 1MGrok 4.5: 45, $3 per 1MClaude Sonnet 5: 45, $4 per 1MClaude Fable 5: 67, $20 per 1MQwen3.7 Plus: 40, $0.7 per 1MClaude Opus 4.8: 49, $10 per 1MGPT-5.5: 61, $11.2 per 1MQwen3.6 27B: 32, $1.35 per 1MKimi K2.6: 34, $1.71 per 1MGrok 4.3: 29, $1.56 per 1MClaude Opus 4.7: 54, $10 per 1MMuse Spark 1.1: 64, $2 per 1MGLM-5V-Turbo: 37, $1.9 per 1MGPT-5.4 mini: 36, $1.69 per 1MGPT-5.4: 54, $5.62 per 1MQwen3.5 27B: 35, $0.82 per 1MGemini 3.1 Pro Preview: 69, $4.5 per 1MClaude Sonnet 4.6: 42, $6 per 1MMistral Large 3: 20, $0.75 per 1MGemini 3 Flash Preview: 54, $1.12 per 1MGemini 3.5 Flash: 67, $3.38 per 1MGemini 3.5 FlashPrice per 1M tokens →↑ Visual reasoning$0.1$1$1020406080GPT-6.1 Terra: 79, $2 per 1MMuse Spark Open: 58, $2 per 1MClaude Haiku 5.5: 59, $0.2 per 1MMistral Large 4: 42, $2.06 per 1MGemini 4 Argon: 76, $4 per 1MGPT-6.1 Sol: 79, $4 per 1MClaude Sonnet 5.5: 71, $4 per 1MGPT-6 Sol: 66, $4 per 1MGPT-6 Luna: 59, $0.2 per 1MClaude Opus 5.5: 76, $8 per 1MGrok 4.7: 56, $3 per 1MDeepSeek V4.1 Flash: 54, $0.21 per 1MGPT-6 Astra: 78, $20 per 1MMuse Spark 1.3: 66, $2 per 1MGemini 3.8 Flash: 66, $1.5 per 1MClaude Fable 5.1: 68, $20 per 1MQwen3.8 Flash: 46, $0.23 per 1MGLM-5.3-Flash: 41, $0.24 per 1MDeepSeek V4 Flash Vision Exp: 22, $0.32 per 1MQwen3.8 27B: 29, $1.12 per 1MGemini 3.7 Flash: 67, $1.5 per 1MGrok 4.6: 51, $3 per 1MMuse Glimmer 30B: 43, $0.52 per 1MMuse Spark 1.2: 56, $2 per 1MQwen3.8 Max: 66, $3 per 1MClaude Opus 5: 60, $10 per 1MGemini 3.6 Flash: 57, $1.5 per 1MGemini 3.5 Flash Lite: 41, $0.85 per 1MKimi K3: 49, $6 per 1MQwen3.7 Flash: 30, $0.055 per 1MGPT-5.6 Terra: 55, $4.5 per 1MGPT-5.6 Sol: 66, $8 per 1MGPT-5.6 Luna: 51, $0.45 per 1MGrok 4.5: 45, $3 per 1MClaude Sonnet 5: 45, $4 per 1MClaude Fable 5: 67, $20 per 1MQwen3.7 Plus: 40, $0.7 per 1MClaude Opus 4.8: 49, $10 per 1MGPT-5.5: 61, $11.2 per 1MQwen3.6 27B: 32, $1.35 per 1MKimi K2.6: 34, $1.71 per 1MGrok 4.3: 29, $1.56 per 1MClaude Opus 4.7: 54, $10 per 1MMuse Spark 1.1: 64, $2 per 1MGLM-5V-Turbo: 37, $1.9 per 1MGPT-5.4 mini: 36, $1.69 per 1MGPT-5.4: 54, $5.62 per 1MQwen3.5 27B: 35, $0.82 per 1MGemini 3.1 Pro Preview: 69, $4.5 per 1MClaude Sonnet 4.6: 42, $6 per 1MMistral Large 3: 20, $0.75 per 1MGemini 3 Flash Preview: 54, $1.12 per 1MGemini 3.5 Flash: 67, $3.38 per 1MGemini 3.5 FlashPrice per 1M tokens →↑ Visual reasoning

Google $1.5 / $9 · OpenRouter $1.5 / $9 · Cached input $0.15

Since Gemini 3 Flash Preview

8 releases in the Gemini line.

Gemini 2.0 FlashDec 11, 2024160dGemini 2.5 Flash Preview 05-20May 20, 202528dGemini 2.5 FlashJun 17, 2025183dGemini 3 Flash PreviewDec 17, 2025153dGemini 3.5 FlashMay 19, 202663dGemini 3.6 FlashJul 21, 202623dGemini 3.7 FlashAug 13, 2026
Released153 days later
Price in / out$1.5 / $9+200%
Context1.05Msame
Coding47−2.2
Writing56+6.8
Reading & documents56−9.3
Reasoning64+6.0
Math54+4.0
Science65+1.5
Visual reasoning67+13.3

A newer release is out: Gemini 3.8 Flash.

ECI STANDING#39 of 270
SCORE154.88152.91–157.41 CI
FRONTIER LAGLaunched 5.2 months behindScores about what GPT-5.2 Pro did in December 2025.
SCORES LIKE THESE MODELSMuse Spark OpenMeta AI · Oct 2026155.11
Qwen3.8 Max (0902)Alibaba · Sep 2026155.08
Muse Spark 1.2Meta AI · Aug 2026155.05DeepSeek V4 Flash 0731DeepSeek · Jul 2026154.76
Full Epoch Capabilities Index ranking ↗

Epoch benchmark boards

Best result on each board it has run. The tick marks the leader.

SimpleBench
77%#5 of 89
SWE-Bench verified
79%#3 of 32
GPQA diamond
93%#22 of 234
SimpleQA Verified
66%#12 of 85
DTBench
95%#26 of 179
OTIS Mock AIME 2024-2025
96%#42 of 216
ARC-AGI-2
72%#21 of 103
ARC-AGI
92%#23 of 107
9 more boards
WeirdML
63%#31 of 141
Mystery Game Puzzles
32%#28 of 88
Chess Puzzles
45%#30 of 93
Surface Evolver Bench
58%#11 of 26
FrontierMath-Tiers-1-3-v2-Private
63%#40 of 90
ProofBench
31%#43 of 81
FrontierMath-Tier-4-v2-Private
27%#44 of 72
APEX-Agents
28%#47 of 51
EBR-bench
5%#26 of 28

Google's past year

37 release days in the past year, about one every 10 days. The next is expected around Nov 3, 2026. Forecasts ↗

NovDecJanFebMarAprMayJunJulAugSepOctNov

Specifications

RELEASE DATEMay 19, 2026
WEIGHTSClosed weights
CONTEXT WINDOW1.05M
MAX OUTPUT65.5K
REASONINGNot listed
TOOL CALLINGNot listed
INPUTStext, image/png, image/jpeg, image/webp, image/heic, image/heif, video/mp4, video/mpeg, video/mov, video/avi, video/x-flv, video/mpg, video/webm, video/wmv, video/3gpp, audio/wav, audio/mp3, audio/aiff, audio/aac, audio/ogg, audio/flac, audio/mpeg, audio/m4a, audio/l16, audio/opus, audio/alaw, audio/mulaw, audio/webm, application/pdf
OUTPUTStext

Models.dev record ↗

Specifications and prices from Models.dev; availability and limits may differ by provider. Benchmarks from Epoch AI (CC BY 4.0), LiveBench and the other sources named above.