A · closed · Gemini

Gemini 3.5 Flash-Lite

Google

B · closed · Gemini

Gemini 3 Flash

Google

Replace A

Replace B

2:7

overall capability wins (1 ties across 10 scored). Gemini 3 Flash leads overall.

Reasoning0:5
Coding1:0
Arena0:1
Tool use0:1
Pricing3:0
Speed2:0
Specs1:0

Verdict

Gemini 3.5 Flash-Lite leads coding (widest gap: +8.7 on WebDev Elo); Gemini 3 Flash edges reasoning & knowledge; Gemini 3 Flash ranks higher on LMArena (+18 Elo).

Gemini 3.5 Flash-Lite leads coding (widest gap: +8.7 on WebDev Elo)
Gemini 3 Flash edges reasoning & knowledge
Gemini 3 Flash ranks higher on LMArena (+18 Elo)
Gemini 3.5 Flash-Lite offers larger context (1.0M vs 1M)
Gemini 3.5 Flash-Lite is ~4.0x cheaper on a blended token basis than Gemini 3 Flash

Cost for 1M in + 250K out

Illustrative chat workload at primary-provider list prices.

Gemini 3.5 Flash-Lite: $0.93
Gemini 3 Flash: $3.75

Gemini 3.5 Flash-Lite is about 4.1x cheaper on this workload.

Price per 1M tokens

Full comparison

Organization

Google

Family

Gemini

License

Proprietary

Open weights

Release

Jul 21, 2026

May 19, 2026

Knowledge cutoff

API / provider

Google

Modalities

text, image, audio, video → text

Specs

Context window

1.0M

A +49K

Max output

66K

tie

Parameters

—

Pricing

Input $/1M

$0.30

$1.50

A +$1.20

Output $/1M

$2.50

A +$6.50

Blended $/1M (3∶1)

$0.85

$3.38

A +$2.52

Speed

tok/s

220

200

A +20

TTFT (s)

0.18

0.2

A +0.02

Reasoning

MMLU-Pro

—

GPQA Diamond

83.8

89.8

B +6.0 pts

Humanity's Last Exam

18.8

36.6

B +17.8 pts

AIME 2025

—

MATH-500

—

Humanity's Last Exam (with tools)

—

AA-Omniscience Accuracy

29.5

53.4

B +23.9 pts

AA-LCR v1.1

76.0

78.0

B +2.0 pts

CritPt

—

8.6

—

MMMU-Pro

79.0

79.9

B +0.9 pts

IFBench

—

78.0

—

Chartography

—

Chartography (With Tools)

—

Coding

SWE-bench Verified

—

SWE-bench Pro

—

34.6

—

SWE-bench Multilingual

—

72.7

—

LiveCodeBench

—

Terminal-Bench 2.1

53.6

—

Aider Polyglot

—

Terminal-Bench 3

—

BigCodeBench

—

SciCode

41.3

—

CursorBench

—

SWE-Rebench

—

NL2Repo-Bench

—

DeepSWE

—

WebDev Arena

1,447

1,438

A +9 Elo

Terminal-Bench 4.0

—

CursorBench 4.0

—

FrontierCode v1.1 (Main)

—

Arena

LMArena Elo

1,456

1,474

B +18 Elo

Metric	Gemini 3.5 Flash-Lite	Gemini 3 Flash	Delta
Identity
Organization	Google	Google	—
Family	Gemini	Gemini	—
License	Proprietary	Proprietary	—
Open weights	No	No	—
Release	Jul 21, 2026	May 19, 2026	—
Knowledge cutoff	-	-	—
API / provider	Google	Google	—
Modalities	text, image, audio, video → text	text, image, audio, video → text	—
Specs
Context window	1.0M	1M	A +49K
Max output	66K	66K	tie
Parameters	—	—	—
Pricing
Input $/1M	$0.30	$1.50	A +$1.20
Output $/1M	$2.50	$9	A +$6.50
Blended $/1M (3∶1)	$0.85	$3.38	A +$2.52
Speed
tok/s	220	200	A +20
TTFT (s)	0.18	0.2	A +0.02
Reasoning
MMLU-Pro	—	—	—
GPQA Diamond	83.8	89.8	B +6.0 pts
Humanity's Last Exam	18.8	36.6	B +17.8 pts
AIME 2025	—	—	—
MATH-500	—	—	—
Humanity's Last Exam (with tools)	—	—	—
AA-Omniscience Accuracy	29.5	53.4	B +23.9 pts
AA-LCR v1.1	76.0	78.0	B +2.0 pts
CritPt	—	8.6	—
MMMU-Pro	79.0	79.9	B +0.9 pts
IFBench	—	78.0	—
Chartography	—	—	—
Chartography (With Tools)	—	—	—
Coding
SWE-bench Verified	—	—	—
SWE-bench Pro	—	34.6	—
SWE-bench Multilingual	—	72.7	—
LiveCodeBench	—	—	—
Terminal-Bench 2.1	53.6	—	—
Aider Polyglot	—	—	—
Terminal-Bench 3	—	—	—
BigCodeBench	—	—	—
SciCode	41.3	—	—
CursorBench	—	—	—
SWE-Rebench	—	—	—
NL2Repo-Bench	—	—	—
DeepSWE	—	—	—
WebDev Arena	1,447	1,438	A +9 Elo
Terminal-Bench 4.0	—	—	—
CursorBench 4.0	—	—	—
FrontierCode v1.1 (Main)	—	—	—
Arena
LMArena Elo	1,456	1,474	B +18 Elo