Qwen3-Embedding-0.6B: benchmark results
Provider: Alibaba. Access: Open.
Unified ELO 1483 ± 20, rank #804 of 1607 rated models, from 22 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BTZSC - Emotion | 43.1 | Macro-F1 (%) | 85.3 |
| SkMTEB - Clustering | 44.28 | Mean V-measure (%) over the five clustering datasets; higher | 85.2 |
| BTZSC - Intent | 56.28 | Macro-F1 (%) | 73.5 |
| BTZSC | 57.97 | Macro-F1 (%) | 61.8 |
| BTZSC - Topic | 48.56 | Macro-F1 (%) | 55.9 |
| SkillRet | 61.94 | NDCG@10 (0-100; first-stage dense retrieval of the gold skil | 52.9 |
| SkMTEB - Classification | 62.84 | Mean accuracy (%) over the seven classification datasets; hi | 51.9 |
| SABER-Math - Geometry | 62.9 | nDCG@10 (0-100) with exponential gain, printed 0 to 1 and sc | 50 |
| SkMTEB | 70.53 | Mean score (%) across the 31 Slovak MTEB datasets over seven | 48.1 |
| SABER-Math - Statement-Statement | 56.6 | Overall nDCG@10 (0-100) with exponential gain, printed 0 to | 47.6 |
| SkMTEB - Pair Classification | 64.46 | Mean score (%) over the three pair-classification datasets; | 44.4 |
| SkMTEB - STS | 83.14 | Mean score (%) over the two semantic-textual-similarity data | 44.4 |
Interactive version: theaggregate.ai/model?slug=qwen3-embedding-0-6b · How It Works · Data refreshed daily, snapshot 2026-09-29.