language benchmark
AlignBench Leaderboard
AlignBench evaluates Chinese alignment of Large Language Models across eight categories using 683 real-scenario rooted queries with human-verified references.
Updated Sep 5, 2026
Models4
Model coverage4
MetricScore
EvidenceB
AlignBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score81.60% | Percentile100.00% | Participants4 | EvidenceC | Evaluated |
| Rank02 | ModelDE | Score80.40% | Percentile66.67% | Participants4 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score73.30% | Percentile33.33% | Participants4 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score72.10% | Percentile0.00% | Participants4 | EvidenceC | Evaluated |
AlignBench Highlights
The leading models and scores on this benchmark.
AlignBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for AlignBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is AlignBench?
What AlignBench measures and how its scores work.
AlignBench is a benchmark for evaluating Chinese alignment of Large Language Models across Fundamental Language Ability, Advanced Chinese Understanding, Open-ended Questions, Writing Ability, Logical Reasoning, Mathematics, Task-oriented Role Play, and Professional Knowledge.
It measures performance in these eight categories using a rule-calibrated multi-dimensional LLM-as-Judge approach with Chain-of-Thought for evaluation.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- AlignBench
- Modality
- text
- Primary category
- language
- 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 AlignBench.
Which model scores highest on AlignBench?
Qwen2.5 72B Instruct is currently ranked first with 81.60%.
What are the top three models on AlignBench?
The current leaders are Qwen2.5 72B Instruct (81.60%), DeepSeek-V2.5 (80.40%), and Qwen2.5 7B Instruct (73.30%).
Which AlignBench model has the lowest official input price?
Qwen2.5 7B Instruct has the lowest matched official input price at $0.18 input / $0.70 output per 1M tokens.
Which models are fastest among AlignBench results?
The fastest matched records are DeepSeek-V2.5 (100.00 tok/s via DeepSeek) and Qwen2.5 72B Instruct (10.00 tok/s via DeepInfra).
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.