The LLM benchmark aggregator.

Grok 4.3

SpaceXAI

Strongest in Science (#22 of 122), weakest in Safety (#253 of 337). Above par in 9 of 28 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 5 and GPT-5.5 and ahead of GLM-4.6 and Gemini 2.5 Flash Lite.

§ 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.

Knowledge54.8−8.7#56/1381/2

Academic knowledge55.8−8.1#52/1231/1

Human preference61.1−6.4#72/3421/1

Professional49.5−12.1#88/1681/1

Legal52.8−14.3#56/1511/1

Finance48.1−14.8#100/1511/1

Medical45.8−21#109/1401/1

Agents50.9−17.2#109/2682/5

Knowledge work51.1−22.5#74/1782/2

Reasoning43.8−23.1#135/1782/4

Science60.7−2.8#22/1221/1

Mathematics42.8−22.9#107/1401/2

Reasoning32.1−37#138/1401/3

Coding37.7−31.4#162/1652/5

Code generation34.6−35.4#72/771/2

Agentic coding38.5−30.2#148/1572/4

Core abilities35−32.4#203/2041/3

Instruction following41−32.7#44/571/1

Language39.3−32.2#48/571/1

Data analysis26−40.1#54/571/1

General intelligence34−36.1#198/2041/3

Safety47.1−14.4#253/3371/3

Secure code56−10.5#109/2741/1

Harm refusal48.1−13.5#186/3001/2

Fairness44−26#219/3001/2

Toxicity avoidance46.7−12.3#222/2721/1

Jailbreak resistance38.6−26.3#225/2721/1

§ 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 Grok 4.3, left for the other.

Ling 3.0 Flashwon 14 · lost 4

Mistral Medium 3.5won 13 · lost 4

Mistral Largewon 12 · lost 5

Mistral Smallwon 13 · lost 6

Grok 4.1 Fastwon 14 · lost 7

Claude 3.5 Haikuwon 13 · lost 7

Llama 4 Scout Instructwon 14 · lost 8

Command Awon 14 · lost 8

Gemini 1.5 Prowon 11 · lost 7

Llama 4 Maverick Instructwon 14 · lost 10

Mistral Large 3won 11 · lost 8

GPT OSS 20Bwon 12 · lost 9

DeepSeek V3won 14 · lost 11

GPT-4.1 Nanowon 13 · lost 11

Gemini 2.5 Flash Litewon 9 · lost 8

GLM-4.6won 12 · lost 11

Grok 3won 10 · lost 10

Gemini 2.0 Flashwon 11 · lost 11

GPT-4o Miniwon 12 · lost 12

DeepSeek R1won 12 · lost 12

GPT-5.4 Nanowon 11 · lost 12

Grok 4 Fastwon 8 · lost 9

DeepSeek V3.2won 11 · lost 13

Gemini 3.5 Flash Litewon 10 · lost 12

Gemini 3.1 Flash Litewon 8 · lost 10

O3 Miniwon 10 · lost 13

Grok 4won 10 · lost 13

Kimi K2won 10 · lost 13

GPT-5 Nanowon 9 · lost 12

Grok Build 0.1won 8 · lost 11

GLM-4.5won 8 · lost 11

GPT OSS 120Bwon 10 · lost 14

Claude Haiku 4.5won 10 · lost 14

MiniMax M3won 11 · lost 17

GPT-4owon 8 · lost 13

MiniMax M2.7won 8 · lost 13

GPT-5.4 Miniwon 8 · lost 14

Claude 3.5 Sonnetwon 7 · lost 13

Mimo V2.5 Prowon 8 · lost 15

Claude 3.7 Sonnetwon 7 · lost 14

Gemma 4 31Bwon 7 · lost 14

GPT-4.1 Miniwon 8 · lost 16

Claude Sonnet 4won 8 · lost 17

Gemini 2.5 Prowon 8 · lost 17

Qwen3.8 27Bwon 7 · lost 15

Nemotron 3 Ultra 550B A55Bwon 6 · lost 13

O1won 6 · lost 14

GPT-4.1won 6 · lost 17

Qwen3 235B A22B Instructwon 6 · lost 17

GPT-5 Miniwon 6 · lost 18

O4 Miniwon 6 · lost 18

Gemini 2.5 Flashwon 6 · lost 18

DeepSeek V4 Flashwon 7 · lost 21

GLM-5.3 Flashwon 7 · lost 21

Inklingwon 6 · lost 22

O3won 5 · lost 19

Kimi K2 Thinkingwon 4 · lost 16

Qwen3.6 Pluswon 4 · lost 18

GLM-5.2won 5 · lost 23

Gemini 3.6 Flashwon 5 · lost 23

Claude Opus 4won 4 · lost 19

GPT-6 Lunawon 4 · lost 21

Kimi K2.7 Codewon 3 · lost 16

Kimi K2.6won 4 · lost 24

Qwen3.7 Maxwon 4 · lost 24

Gemini 3.5 Flashwon 4 · lost 24

GLM-5.1won 3 · lost 19

Gemini 3.1 Prowon 3 · lost 20

GPT-5won 3 · lost 21

Claude Sonnet 4.5won 3 · lost 21

GLM-5won 2 · lost 15

GPT-5.2won 3 · lost 24

GPT-5.6 Lunawon 3 · lost 25

DeepSeek V4.1 Flashwon 2 · lost 19

Claude Opus 4.5won 2 · lost 20

Kimi K2.5won 2 · lost 22

Claude Opus 4.6won 2 · lost 22

Grok 4.7won 2 · lost 23

GPT-6 Solwon 2 · lost 23

Claude Sonnet 4.6won 2 · lost 24

GPT-5.6 Terrawon 2 · lost 26

Gemini 3.8 Flashwon 2 · lost 26

Mimo V2.6 Prowon 1 · lost 18

Gemini 3 Prowon 1 · lost 19

GPT-5.1won 1 · lost 21

GLM-5.3won 1 · lost 21

Claude Sonnet 5won 1 · lost 27

DeepSeek V4 Prowon 1 · lost 27

Grok 4.5won 1 · lost 27

GPT-5.4won 1 · lost 27

Claude Opus 4.8won 1 · lost 27

GPT-5.6 Solwon 1 · lost 27

Muse Spark 1.1won 1 · lost 27

Claude Opus 4.7won 1 · lost 27

Muse Spark 1.3won 0 · lost 18

GPT-6 Astrawon 0 · lost 20

Muse Spark 1.2won 0 · lost 21

Gemini 3.7 Flashwon 0 · lost 22

Kimi K3won 0 · lost 22

Claude Fable 5.1won 0 · lost 22

Claude Fable 5won 0 · lost 22

Claude Opus 5.5won 0 · lost 25

Grok 4.6won 0 · lost 28

Qwen3.8 Maxwon 0 · lost 28

GPT-5.5won 0 · lost 28

Claude Opus 5won 0 · lost 28

§ 3 · Sources

Where the numbers come from

6 publications, 32 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26

LMArena Text 1442

GDPval-AA ↗1 measureread 2026-09-26

GDPval-AA 923

AA-Briefcase ↗1 measureread 2026-09-26

AA-Briefcase 761

LiveBench ↗8 measuresread 2026-09-26

LiveBench 62.3LiveBench · Reasoning 70.8LiveBench · Coding 69.9LiveBench · Agentic Coding 18.5LiveBench · Mathematics 84.3LiveBench · Data Analysis 55.8LiveBench · Language 73.6LiveBench · Instruction Following 62.8

Vals.ai ↗15 measuresread 2026-09-26

Vals · Legal Research Bench 15.38%Vals · CaseLaw 79.31%Vals · LegalBench 84.46%Vals · Harvey Legal Agent Benchmark 0.42%Vals · Finance Agent 37.73%Vals · CorpFin 68.53%Vals · TaxEval 70.81%Vals · MortgageTax 48.25%Vals · MedCode 38.07%Vals · MedScribe 74.4%Vals · SWE-bench Verified 71.4% (Mini-SWE-agent)Vals · Vibe Code Bench 19.4% (OpenHands)Vals · Code Migration 6.79%Vals · GPQA Diamond 91.41%Vals · MMLU Pro 85.84%

Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-09-26

Enkrypt · Jailbreak risk 22.4%Enkrypt · Harmful content risk 5%Enkrypt · CBRN risk 34.8%Enkrypt · Toxicity risk 8.6%Enkrypt · Bias risk 87.1%Enkrypt · Insecure code risk 21.3%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=grok-4-3)](https://publicai.io/model-index/m/grok-4-3)