reasoning benchmark
HumanEval-Mul Leaderboard
HumanEval-Mul is a multilingual variant of the HumanEval benchmark consisting of 164 original programming problems.
Updated Sep 5, 2026
Models2
Model coverage2
MetricScore
EvidenceB
HumanEval-Mul Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelDE | Score82.60% | Percentile100.00% | Participants2 | EvidenceC | Evaluated |
| Rank02 | ModelDE | Score73.80% | Percentile0.00% | Participants2 | EvidenceC | Evaluated |
HumanEval-Mul Highlights
The leading models and scores on this benchmark.
HumanEval-Mul Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for HumanEval-Mul
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is HumanEval-Mul?
What HumanEval-Mul measures and how its scores work.
HumanEval-Mul is a benchmark for synthesizing programs from docstrings.
It measures functional correctness, language comprehension, algorithms, and simple mathematics, with Score reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- HumanEval-Mul
- Modality
- text
- Primary category
- reasoning
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about HumanEval-Mul.
Which model scores highest on HumanEval-Mul?
DeepSeek-V3 is currently ranked first with 82.60%.
What are the top three models on HumanEval-Mul?
The current leaders are DeepSeek-V3 (82.60%) and DeepSeek-V2.5 (73.80%).
Which HumanEval-Mul model has the lowest official input price?
No matched official input price is currently available.
Which models are fastest among HumanEval-Mul results?
The fastest matched records are DeepSeek-V3 (100.00 tok/s via DeepSeek) and DeepSeek-V2.5 (100.00 tok/s via DeepSeek).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.