LLMcompare

Search

Search for a command to run...

A · closed · Gemini

Gemini 2.5 Pro

Google

B · closed · Gemini

Gemini 3.5 Flash-Lite

Google

4:7

overall capability wins (1 ties across 12 scored). Gemini 3.5 Flash-Lite leads overall.

  • Reasoning3:2
  • Coding1:2
  • Arena0:1
  • Tool use0:1
  • Composite indices0:2
  • Pricing0:3
  • Speed0:2
  • Specs0:1

Verdict

Gemini 3.5 Flash-Lite leads coding (widest gap: +220.9 on WebDev Elo); Gemini 2.5 Pro edges reasoning & knowledge; Gemini 3.5 Flash-Lite ranks higher on LMArena (+10 Elo).

  • Gemini 3.5 Flash-Lite leads coding (widest gap: +220.9 on WebDev Elo)
  • Gemini 2.5 Pro edges reasoning & knowledge
  • Gemini 3.5 Flash-Lite ranks higher on LMArena (+10 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 2.5 Pro

Cost for 1M in + 250K out

Illustrative chat workload at primary-provider list prices.

Gemini 2.5 Pro
$3.75
Gemini 3.5 Flash-Lite
$0.93

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

Price per 1M tokens

Full comparison

Organization

Google

Google

Family

Gemini

Gemini

License

Proprietary

Proprietary

Open weights

No

No

Release

Mar 25, 2025

Jul 21, 2026

Knowledge cutoff

-

-

API / provider

Google

Google

Modalities

text, image, audio, video → text

text, image, audio, video → text

Specs

Context window

1M

1.0M

B +49K

Max output

66K

66K

tie

Parameters

—

—

—

Pricing

Input $/1M

$1.25

$0.30

B +$0.95

Output $/1M

$10

$2.50

B +$7.50

Blended $/1M (3∶1)

$3.44

$0.85

B +$2.59

Speed

tok/s

100

220

B +120

TTFT (s)

0.45

0.18

B +0.27

Reasoning

MMLU-Pro

86.2

—

—

GPQA Diamond

84.4

83.8

A +0.6 pts

Humanity's Last Exam

22.5

18.8

A +3.7 pts

AIME 2025

87.7

—

—

MATH-500

96.7

—

—

Humanity's Last Exam (with tools)

—

—

—

AA-Omniscience Accuracy

39.1

29.5

A +9.6 pts

AA-LCR v1.1

69.0

76.0

B +7.0 pts

CritPt

2.6

—

—

MMMU-Pro

74.9

79.0

B +4.1 pts

IFBench

48.7

—

—

Chartography

—

—

—

Chartography (With Tools)

—

—

—

Coding

SWE-bench Verified

53.6

—

—

SWE-bench Pro

—

—

—

SWE-bench Multilingual

—

—

—

LiveCodeBench

80.1

—

—

Terminal-Bench 2.1

28.5

53.6

B +25.1 pts

Aider Polyglot

79.1

—

—

Terminal-Bench 3

—

—

—

BigCodeBench

—

—

—

SciCode

46.3

41.3

A +5.0 pts

CursorBench

—

—

—

SWE-Rebench

23.5

—

—

NL2Repo-Bench

—

—

—

DeepSWE

—

—

—

WebDev Arena

1,226

1,447

B +221 Elo

Terminal-Bench 4.0

—

—

—

CursorBench 4.0

—

—

—

FrontierCode v1.1 (Main)

—

—

—

Arena

LMArena Elo

1,446

1,456

B +10 Elo

Gemini 2.5 Pro

Context
1M / 66K out
Parameters
—
Price
$1.25 / $10
Speed
100 tok/s · 0.45s TTFT
Modalities
text, image, audio, video → text
License
Proprietary

Previous-gen Google Pro model; still a capable long-context multimodal option at competitive rates.

Gemini 3.5 Flash-Lite

Context
1.0M / 66K out
Parameters
—
Price
$0.30 / $2.50
Speed
220 tok/s · 0.18s TTFT
Modalities
text, image, audio, video → text
License
Proprietary

GA cost-efficient Flash-Lite for high-volume agentic subtasks, translation, and simple data processing.

Benchmark charts

Winner bars are emphasized. Per-benchmark deltas sit above each chart.

Reasoning & knowledge

GPQA: A +0.6HLE: A +3.7AA-LCR: B +7.0Omniscience Acc.: A +9.6MMMU-Pro: B +4.1

Coding

TermBench: B +25.1SciCode: A +5.0WebDev Elo: B +221

Tool use & function calling

τ³-Banking: B +7.8

Arena

Arena Elo: B +10

Composite indices

AA Index: B +10.9AA Index v4.3: B +6.0

Pick a different pair · Back to catalog