Gemini 3.5 Flash

May 2026
Older versionReplace withGemini 3.8 FlashNewer versionClaude Opus 5.5Cheaper, as goodGPT-6.1 SolBetterMiMo v2.6 ProOpen weights
Gemini 3.5 Flash
Compare in
on
$3.69per task?
$1.50 / $9per 1M in / outcache $0.15
208 tok/sfirst token in 16 s?
1Mcontext

Rankingat high effort?

Retired models are left out of the current rankings.

Alternatives?

ModelScoreCostPer taskWhat changes
CheaperClaude Opus 5.5xhigh+72%-6%$3.46No forced tool use+5BetterMiMo v2.6 FlashthinkingOpen+16%-98%$0.062Pass back reasoning_content+3OpenMiMo v2.6 Prothinking+42%-96%$0.13Pass back reasoning_content+3CheaperGPT-6 Astramax+62%-12%$3.26Monitoring can return HTTP 403+2CheaperGemini 4 Argonhigh+61%-46%*$1.99Synchronous output stops at 65,536+2

History

7 versions
10203040Jul 25Oct 25Jan 26Apr 26Jul 26Release dateAA Intelligence Index 4.3.22.53.53.63.73.8
10203040Jul 25Jan 26Jul 26Release dateAA Intelligence Index 4.3.22.53.53.63.73.8

Benchmarks

all categories
Chat & Knowledge
Humanity's Last Exam42.7
Arena Text · overall1,477
Arena Text · Chinese1,525.4
Arena Text · English1,480.5
Arena Text · French1,489.5
Arena Text · German1,492.2
Arena Text · instruction following1,467.1
Arena Text · Japanese1,486.2
Arena Text · Korean1,445.4
Arena Text · non-English1,475.4
Arena Text · Polish1,516.5
Arena Text · Russian1,492.4
Arena Text · Spanish1,479.6
AA Omniscience Index21.2
SimpleQA Verified v166.2
CritPt (physics)13.1
AA Intelligence Index 4.3.232.6
AA Omniscience · Accuracy51.4
AA Omniscience · Hallucination Rate62.2
AA-LCR (long context) 1.173.3
GDP.pdf · AA all-pass19.8
MLCR-AA18.3
ARC-AGI-1 · public evaluation96
ARC-AGI-2 · public evaluation72.1
ARC-AGI-1 · semi-private92.5
ARC-AGI-2 · semi-private72.1
FrontierMath39
FrontierMath Tier 414.6
FrontierMath Tiers 1–3 (v2)62.8
FrontierMath Tier 4 (v2)26.8
Chess Puzzles45
Mystery Game Puzzles32
Arena Text · overall · full cohort1,482
GPQA Diamond92.2
GPQA Diamond · Epoch92.8
OTIS Mock AIME 2024 202595.6
Coding
Terminal-Bench 4.06.6
DeepSWE 1.136.1
WebDev Arena1,499
SciCode53.9
Terminal-Bench Hard (version unstated)40.9
SWE-bench Pro (public)lab55.1
WebDev Arena · full cohort1,499.4
Terminal-Bench 2.178.7
SWE-bench Verified · Epoch79.3
Agents & Tool Use
AutomationBench-AA42.1
GDPval-AA (work tasks) 2.135.1
Toolathlon-Verified67.3
τ³-Banking · AA32.2
AA Analyst Agent45
ITBench SRE40.3
APEX-Agents47.1
EnterpriseOps-Gym-AA50.1
AA-Briefcase 1.1 · Elo874.3
EBR Bench4.8
τ²-bench · AA95.3
Writing & Language
Arena Text · creative writing1,464
IFBench (instructions)76.3
Arena Text · creative writing · full cohort1,468.1
Vision & Documents
MMMU-Pro (vision)84.3
Arena Document · overall1,461.4
Arena Vision · overall1,308.3

FAQ

How much does Gemini 3.5 Flash cost?›

$1.50 per 1M input tokens and $9 per 1M output tokens; cached input $0.15 per 1M. A typical benchmark task costs about $3.69 at high effort.

Is Gemini 3.5 Flash free to use?›

There is no free API; the cheapest paid price is $1.50 per 1M input tokens.

How can I use Gemini 3.5 Flash?›

Through the API of DeepInfra, Google, Google AI Studio.

What is the context window of Gemini 3.5 Flash?›

1M tokens, with up to 66K output tokens.

When was Gemini 3.5 Flash released? Is there a newer version?›

Released 2026-05-19. It has been superseded by Gemini 3.6 Flash.

Is Gemini 3.5 Flash open source?›

No, Gemini 3.5 Flash is a closed model available through an API. The closest open-weights option here is MiMo v2.6 Pro.

Can Gemini 3.5 Flash read images?›

It accepts text and images, PDF files, audio, video as input.

What is Gemini 3.5 Flash?›

Gemini 3.5 Flash is a closed language model from Google, released May 2026, the successor to Gemini 3 Flash Preview.

What are the best alternatives to Gemini 3.5 Flash?›

Claude Opus 5.5 (cheaper), MiMo v2.6 Flash (better), MiMo v2.6 Pro (open weights).