live
weekly refresh
basedagi.org
benchmark evidence

Global-MMLU-Lite Japanese

Artificial Analysis Global-MMLU-Lite accuracy for Japanese.

winner on Global-MMLU-Lite Japanese
direct benchmark result, not a broad vertical composite | source row dated 2000-01-01
scored on 2000-01-01 · stale source data (9646d)
latest mapped results | top 20
#ModelScoreEvidenceTested
1Anthropic: Claude Opus 4.6
Anthropic
93.1
model-only
independent_benchmark
2000-01-01
2Anthropic: Claude Opus 4.5
Anthropic
92.6
model-only
independent_benchmark
2000-01-01
3Anthropic: Claude Sonnet 4.6
Anthropic
92.6
model-only
independent_benchmark
2000-01-01
4Google: Gemini 2.5 Pro
Google
91.0
model-only
independent_benchmark
2000-01-01
5Anthropic: Claude Sonnet 4.5
Anthropic
90.7
model-only
independent_benchmark
2000-01-01
6Google: Gemini 2.5 Flash
Google
89.3
model-only
independent_benchmark
2000-01-01
7OpenAI: GPT-5.2
Openai
88.8
model-only
independent_benchmark
2000-01-01
8Z.ai: GLM 5
Z Ai
87.8
model-only
independent_benchmark
2000-01-01
9Anthropic: Claude Haiku 4.5
Anthropic
86.3
model-only
independent_benchmark
2000-01-01
10Meta: Llama 4 Maverick
Meta Llama
85.3
model-only
independent_benchmark
2000-01-01
11DeepSeek: DeepSeek V3.2
Deepseek
83.3
model-only
independent_benchmark
2000-01-01
12Meta: Llama 4 Scout
Meta Llama
80.5
model-only
independent_benchmark
2000-01-01
13MoonshotAI: Kimi K2 Thinking
Moonshotai
74.9
model-only
independent_benchmark
2000-01-01
what this result means

Artificial Analysis Global-MMLU-Lite accuracy for Japanese.

This benchmark contributes direct public evidence. Read its scope before generalizing the result.

A win here is a win on Global-MMLU-Lite Japanese. Broad task pages require independent corroboration before naming a general winner.

source record
category: multilingual
metric: accuracy
matched models: 13
latest source date: 2000-01-01
direction: higher is better
inspect upstream source ->