The LLM benchmark aggregator.

Claude Opus 4.7

Anthropic

Strongest in Harm refusal (#2 of 300, on 1 of its 2 boards), weakest in Fairness (#101 of 300). Above par in 26 of 29 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of DeepSeek V4.1 Flash and Gemini 3.8 Flash.

§ 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 preference66.8−0.7#4/3421/1

Knowledge59−4.5#10/1381/2

Academic knowledge60.8−3.1#8/1231/1

Professional58.4−3.2#11/1681/1

Medical60.1−6.7#8/1401/1

Legal58.8−8.3#17/1511/1

Finance58.2−4.7#17/1511/1

Reasoning57.8−9.1#22/1783/4

Mathematics58.5−7.2#15/1402/2

Science59.8−3.7#26/1221/1

Reasoning56.1−13#36/1402/3

Coding56.6−12.5#26/1652/5

Code generation59.6−10.4#11/771/2

Agentic coding55.8−12.9#35/1572/4

Safety56.5−5#33/3372/3

Harm refusal60.9−0.7#2/3001/2

Jailbreak resistance63.1−1.8#9/2721/1

Toxicity avoidance56.8−2.2#48/2721/1

Factual grounding44.9−25.7#75/1011/1

Secure code57.8−8.7#88/2741/1

Fairness50.9−19.1#101/3001/2

Agents58.8−9.3#39/2682/5

Knowledge work61.5−12.1#33/1782/2

Core abilities55.1−12.3#43/2042/3

General intelligence61−9.1#18/2042/3

Data analysis58.2−7.9#18/571/1

Instruction following47.9−25.8#33/571/1

Language47.4−24.1#34/571/1

§ 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 Claude Opus 4.7, left for the other.

Ling 3.0 Flashwon 18 · lost 0

Gemini 1.5 Prowon 18 · lost 0

Grok 4 Fastwon 18 · lost 0

Mistral Large 3won 20 · lost 0

Gemini 2.0 Flashwon 22 · lost 0

Qwen3.6 Pluswon 22 · lost 0

GPT-4o Miniwon 24 · lost 0

Grok 4.3won 27 · lost 1

GLM-5.2won 27 · lost 1

Claude Sonnet 4.6won 26 · lost 1

DeepSeek V3won 25 · lost 1

Llama 4 Maverick Instructwon 24 · lost 1

DeepSeek V3.2won 24 · lost 1

GPT-4.1 Nanowon 23 · lost 1

GPT-4.1 Miniwon 23 · lost 1

Llama 4 Scout Instructwon 22 · lost 1

Command Awon 22 · lost 1

GPT-5.4 Miniwon 22 · lost 1

Grok 4.1 Fastwon 21 · lost 1

Gemini 3.5 Flash Litewon 21 · lost 1

GLM-5.1won 21 · lost 1

Grok 3won 20 · lost 1

Kimi K2 Thinkingwon 19 · lost 1

Mistral Smallwon 19 · lost 1

Kimi K2.7 Codewon 18 · lost 1

Grok Build 0.1won 18 · lost 1

GLM-4.5won 18 · lost 1

Gemini 3.1 Flash Litewon 18 · lost 1

MiniMax M3won 26 · lost 2

GPT OSS 120Bwon 23 · lost 2

GPT-5.4 Nanowon 22 · lost 2

Kimi K2won 22 · lost 2

GLM-4.6won 22 · lost 2

O3 Miniwon 21 · lost 2

Grok 4won 21 · lost 2

Mimo V2.5 Prowon 21 · lost 2

Gemma 4 31Bwon 20 · lost 2

GPT OSS 20Bwon 19 · lost 2

Claude 3.7 Sonnetwon 19 · lost 2

Kimi K2.6won 26 · lost 3

Nemotron 3 Ultra 550B A55Bwon 17 · lost 2

Gemini 3.6 Flashwon 25 · lost 3

Mistral Largewon 16 · lost 2

Gemini 2.5 Flash Litewon 16 · lost 2

GLM-5won 16 · lost 2

Gemini 2.5 Prowon 23 · lost 3

Gemini 2.5 Flashwon 22 · lost 3

Claude Haiku 4.5won 22 · lost 3

Kimi K2.5won 22 · lost 3

Claude Sonnet 4.5won 22 · lost 3

GPT-4.1won 21 · lost 3

Claude Opus 4.5won 20 · lost 3

GPT-4owon 19 · lost 3

Qwen3.8 27Bwon 19 · lost 3

MiniMax M2.7won 18 · lost 3

GLM-5.3 Flashwon 24 · lost 4

GPT-5 Miniwon 21 · lost 4

Qwen3 235B A22B Instructwon 20 · lost 4

Claude Opus 4won 20 · lost 4

DeepSeek V4 Flashwon 23 · lost 5

Inklingwon 23 · lost 5

Qwen3.7 Maxwon 23 · lost 5

GPT-5 Nanowon 18 · lost 4

Claude 3.5 Haikuwon 16 · lost 4

O1won 16 · lost 4

O4 Miniwon 20 · lost 5

DeepSeek R1won 20 · lost 5

GPT-5won 20 · lost 5

GPT-6 Lunawon 20 · lost 5

Gemini 3.5 Flashwon 22 · lost 6

Claude Sonnet 4won 20 · lost 6

Claude Opus 4.6won 19 · lost 6

O3won 18 · lost 6

GPT-5.2won 21 · lost 7

GPT-5.6 Lunawon 21 · lost 7

GPT-5.1won 17 · lost 6

Claude 3.5 Sonnetwon 14 · lost 6

DeepSeek V4 Prowon 20 · lost 9

Claude Sonnet 5won 19 · lost 9

Muse Spark 1.1won 19 · lost 9

GPT-5.4won 19 · lost 10

Grok 4.5won 18 · lost 10

GLM-5.3won 14 · lost 8

Mimo V2.6 Prowon 12 · lost 7

Qwen3.8 Maxwon 17 · lost 11

Gemini 3.7 Flashwon 13 · lost 9

Gemini 3.1 Prowon 14 · lost 10

Gemini 3 Prowon 12 · lost 9

Muse Spark 1.2won 12 · lost 9

Grok 4.7won 14 · lost 11

Gemini 3.8 Flashwon 15 · lost 13

DeepSeek V4.1 Flashwon 11 · lost 10

Grok 4.6won 14 · lost 14

GPT-5.6 Terrawon 14 · lost 14

GPT-5.6 Solwon 14 · lost 15

GPT-6 Solwon 12 · lost 14

GPT-5.5won 13 · lost 16

Claude Opus 4.8won 12 · lost 16

GPT-6 Astrawon 8 · lost 13

Kimi K3won 8 · lost 14

Claude Opus 5won 9 · lost 19

Claude Opus 5.5won 6 · lost 19

Muse Spark 1.3won 4 · lost 14

Claude Fable 5.1won 2 · lost 20

Claude Fable 5won 0 · lost 22

§ 3 · Sources

Where the numbers come from

9 publications, 36 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26

LMArena Text 1502

GDPval-AA ↗1 measureread 2026-09-26

GDPval-AA 1338

AA-Briefcase ↗1 measureread 2026-09-26

AA-Briefcase 1253

LiveBench ↗8 measuresread 2026-09-26

LiveBench 76.5LiveBench · Reasoning 87.2LiveBench · Coding 82.1LiveBench · Agentic Coding 50.7LiveBench · Mathematics 92.9LiveBench · Data Analysis 78.3LiveBench · Language 77.9LiveBench · Instruction Following 66.7

Kagi LLM Benchmark ↗1 measureread 2026-09-26

Kagi LLM Benchmark 80.7%

SimpleBench ↗1 measureread 2026-09-26

SimpleBench 61.7%

Vals.ai ↗16 measuresread 2026-09-26

Vals · Legal Research Bench 38.46%Vals · CaseLaw 68.38%Vals · LegalBench 85.25%Vals · Harvey Legal Agent Benchmark 6.67%Vals · Finance Agent 51.51%Vals · CorpFin 66.08%Vals · TaxEval 75.27%Vals · MortgageTax 70.27%Vals · MedCode 54.86%Vals · MedScribe 82.95%Vals · SWE-bench Verified 82% (Mini-SWE-agent)Vals · Vibe Code Bench 71% (OpenHands)Vals · Code Migration 43.88%Vals · GPQA Diamond 90.15%Vals · MMLU Pro 89.87%Vals · AIME 96.25%

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

Enkrypt · Jailbreak risk 1.7%Enkrypt · Harmful content risk 0.6%Enkrypt · CBRN risk 3.7%Enkrypt · Toxicity risk 1.5%Enkrypt · Bias risk 76.5%Enkrypt · Insecure code risk 17.8%

Vectara Hallucination Leaderboard ↗1 measureread 2026-09-26

Vectara · Factual consistency 88%

Badge

[![PublicAI Index](https://publicai.io/model-index/badge?model=claude-opus-4-7)](https://publicai.io/model-index/m/claude-opus-4-7)