The LLM benchmark aggregator.

Gemini 3.5 Flash-Lite

Google

Strongest in Instruction following (#34 of 62), weakest in Reasoning (#158 of 195). Above par in 7 of 22 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5.1 and GPT-5.6 Sol and ahead of GPT-5.4 mini and NVIDIA Nemotron 3 Ultra 550B A55B.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Human preference62−6.8#61/3541/1

Knowledge51.5−11#94/1972/3

Document parsing50.8−12.5#46/831/1

Academic knowledge55.8−8.1#53/1231/1

Professional49.7−12.7#101/1941/2

Cybersecurity44.8−19.1#37/471/1

Finance53.8−8.6#67/1621/1

Medical47−19.5#104/1491/1

Legal47−18.9#111/1731/1

Coding44.1−22.1#134/1792/6

Agentic coding43.8−23.1#130/1732/5

Agents48.9−17.5#151/2982/6

Knowledge work48.2−22.5#110/2082/2

Core abilities44.5−20.8#152/1992/4

Instruction following48.6−25.2#34/621/1

Language41.7−24.9#54/621/1

General intelligence42.7−26.7#136/1781/3

Reasoning42.7−22.7#158/1953/5

Science55.4−8.1#56/1221/1

Reasoning43.9−25.1#95/1471/3

Mathematics32.9−30.9#150/1521/2

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Gemini 3.5 Flash-Lite, left for the other.

NVIDIA Nemotron 3.5 Lightning 30B A3Bwon 17 · lost 0

GPT-4.1 nanowon 17 · lost 1

GPT-4o miniwon 16 · lost 1

Mistral Medium 3.5won 16 · lost 1

Llama 4 Scout Instructwon 14 · lost 2

Gemma 4 31Bwon 13 · lost 2

GPT-5.4 nanowon 16 · lost 4

Claude 3.5 Haikuwon 11 · lost 3

Llama 4 Maverick Instructwon 14 · lost 4

Claude Haiku 4.5won 13 · lost 4

Command Awon 12 · lost 4

GPT OSS 120Bwon 13 · lost 5

Gemini 3.1 Flash-Litewon 13 · lost 5

GPT-4owon 11 · lost 5

O3 Miniwon 11 · lost 5

GPT OSS 20Bwon 10 · lost 5

GPT-5 nanowon 10 · lost 5

Mistral Large 3won 12 · lost 6

Gemini 2.0 Flashwon 10 · lost 6

DeepSeek V3won 11 · lost 7

MiniMax M2.7won 9 · lost 6

MiMo V2.5won 9 · lost 6

DeepSeek R1won 10 · lost 7

Gemini 2.5 Flash-Litewon 8 · lost 6

Claude Sonnet 4won 10 · lost 8

Qwen3.6 Pluswon 10 · lost 8

Grok 4.1 Fastwon 8 · lost 7

GPT-5 miniwon 10 · lost 9

NVIDIA Nemotron 3 Ultra 550B A55Bwon 10 · lost 9

GPT-5.4 miniwon 10 · lost 9

MiniMax M2.1won 7 · lost 7

MiniMax M2.5won 8 · lost 8

Grok 4 Fastwon 8 · lost 8

GLM-4.7won 8 · lost 8

DeepSeek V3.2won 9 · lost 9

O4 Miniwon 8 · lost 9

Kimi K2won 8 · lost 9

GLM-4.6won 8 · lost 9

Claude 3.7 Sonnetwon 7 · lost 8

GPT-4.1 miniwon 8 · lost 10

Grok 3 Miniwon 7 · lost 9

Inklingwon 9 · lost 12

Gemini 2.5 Flashwon 7 · lost 10

GPT-4.1won 7 · lost 10

Qwen3 235B A22B Instructwon 7 · lost 10

Qwen3 Maxwon 6 · lost 9

MiMo V2.5 Prowon 6 · lost 9

Grok 4.3won 7 · lost 11

Gemini 2.5 Prowon 7 · lost 11

Grok 3won 6 · lost 10

Grok 4.20won 6 · lost 10

O1won 5 · lost 9

Claude Haiku 5.5won 7 · lost 13

GPT-6 Lunawon 7 · lost 13

Kimi K2.5won 6 · lost 12

Mistral Large 4won 5 · lost 12

Kimi K2 Thinkingwon 4 · lost 10

Step 5won 4 · lost 10

MiniMax M3won 6 · lost 15

Qwen3.8 27Bwon 6 · lost 16

Inkling Smallwon 5 · lost 14

Claude Opus 4won 4 · lost 12

Grok 4won 4 · lost 13

DeepSeek V4 Flashwon 4 · lost 17

GLM-5.1won 3 · lost 13

GPT-5.6 Lunawon 4 · lost 18

O3won 3 · lost 14

Grok 4.7won 3 · lost 15

GPT-5.2won 3 · lost 16

Claude Opus 4.5won 3 · lost 16

Claude Opus 4.6won 3 · lost 16

MiMo V2.6 Flashwon 2 · lost 13

Gemini 3 Flashwon 2 · lost 13

GLM-5won 2 · lost 14

GPT-5won 2 · lost 16

Claude Sonnet 4.5won 2 · lost 16

GPT-6 Astrawon 2 · lost 18

GPT-6.1 Solwon 2 · lost 18

Gemini 3.1 Prowon 2 · lost 18

Claude Opus 5.5won 2 · lost 18

MiMo V2.6 Prowon 1 · lost 15

Kimi K2.6won 1 · lost 17

GPT-6 Solwon 1 · lost 18

Claude Sonnet 4.6won 1 · lost 18

GLM-5.2won 1 · lost 19

Claude Opus 5won 1 · lost 19

Claude Sonnet 5won 1 · lost 20

Gemini 3.6 Flashwon 1 · lost 20

Qwen3.8 Maxwon 1 · lost 20

Gemini 3.7 Flashwon 1 · lost 20

Claude Opus 4.7won 1 · lost 20

GPT-5.6 Terrawon 1 · lost 21

GLM-5.3 Flashwon 1 · lost 21

Gemini 3.8 Flashwon 1 · lost 21

Muse Sparkwon 0 · lost 15

GPT-5.1won 0 · lost 16

Gemini 4 Argonwon 0 · lost 16

Gemini 3 Prowon 0 · lost 17

Muse Spark 1.3won 0 · lost 17

Muse Wonderwon 0 · lost 17

Muse Spark 1.1won 0 · lost 18

Qwen3.7 Maxwon 0 · lost 19

Muse Spark 1.2won 0 · lost 19

Claude Sonnet 5.5won 0 · lost 20

DeepSeek V4.1 Flashwon 0 · lost 20

Gemini 3.5 Flashwon 0 · lost 20

Grok 4.6won 0 · lost 21

DeepSeek V4 Prowon 0 · lost 21

Grok 4.5won 0 · lost 21

GPT-5.4won 0 · lost 21

GPT-5.5won 0 · lost 21

Claude Opus 4.8won 0 · lost 21

GLM-5.3won 0 · lost 21

Kimi K3won 0 · lost 21

Claude Fable 5won 0 · lost 21

GPT-5.6 Solwon 0 · lost 22

Claude Fable 5.1won 0 · lost 22

§ 3 · Sources

Where the numbers come from

8 publications, 33 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-11-10

LMArena Text 1455

Artificial Analysis ↗1 measureread 2026-11-10

Artificial Analysis Intelligence Index 12

GDPval-AA ↗1 measureread 2026-11-10

GDPval-AA 995

AA-Briefcase ↗1 measureread 2026-11-10

AA-Briefcase 651

ARC-AGI-2 ↗1 measureread 2026-11-10

ARC-AGI-2 10.3%

LiveBench 1.0 ↗8 measuresread 2026-11-10

LiveBench 1.0 63.9LiveBench 1.0 · Reasoning 60.2LiveBench 1.0 · Coding 76.1LiveBench 1.0 · Agentic Coding 45.3LiveBench 1.0 · Mathematics 73.7LiveBench 1.0 · Data Analysis 53.2LiveBench 1.0 · Language 71.8LiveBench 1.0 · Instruction Following 67.2

Vals.ai ↗14 measuresread 2026-11-10

Vals · Legal Research Bench 13.94%Vals · LegalBench 84.07%Vals · Finance Agent 47.44%Vals · CorpFin 60.68%Vals · TaxEval 72.61%Vals · MortgageTax 68.68%Vals · MedCode 43.49%Vals · MedScribe 70.89%Vals · CyberBench 59.11%Vals · SWE-bench Verified 75% (Mini-SWE-agent)Vals · Vibe Code Bench 37.16% (OpenHands)Vals · Code Migration 6.15%Vals · GPQA Diamond 83.84%Vals · MMLU Pro 85.84%

ParseBench ↗6 measuresread 2026-11-10

ParseBench · Overall 59.49%ParseBench · Tables 83.35%ParseBench · Charts 19.27%ParseBench · Content faithfulness 86.48%ParseBench · Semantic formatting 54.05%ParseBench · Visual grounding 54.3%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=gemini-3-5-flash-lite)](https://publicai.io/model-index/m/gemini-3-5-flash-lite)