Key Specifications
| Vendor | alibaba |
|---|
| Version | 1.5-72b |
|---|
| Release Date | 2024-02-25 |
|---|
| Context Window | 32768 tokens |
|---|
| Input Modalities | text |
|---|
| Output Modalities | text |
|---|
| License | Qwen License |
|---|
| Documentation | https://qwenlm.github.io/ |
|---|
Benchmark Performance
| Benchmark | Score | Unit | Evaluated At | Notes | Source |
|---|
| MMLU | 81.3 | % | 2024-02-25 | 5-shot | view |
| HUMANEVAL | 65.2 | pass@1 | 2024-02-25 | — | view |
| GSM8K | 80.9 | % | 2024-02-25 | 0-shot CoT | view |
| MATH | 37.6 | % | 2024-02-25 | 0-shot CoT | view |
| BBH | 76.5 | % | 2024-02-25 | 3-shot CoT | view |
| GPQA | 39.8 | % | 2024-02-25 | 0-shot | view |
| IFEVAL | 76.4 | % | 2024-02-25 | prompt_strict | view |
| ARC | 92.6 | % | 2024-02-25 | challenge | view |
| MUSR | 50.8 | % | 2024-02-25 | 0-shot | view |
| WINOGRANDE | 83.1 | % | 2024-02-25 | 0-shot | view |
Compliance
- Data Residency: CN
- SOC2: ✗
- HIPAA: ✗
- GDPR: ✗
- ISO 27001: ✗
Qwen1.5 72B
Modellübersicht
阿里巴巴 Qwen1.5 72B 开源模型, 32K 上下文, 中文理解与生成能力突出, 适合企业部署。
Kernspezifikationen
| Anbieter | Version | Veröffentlichungsdatum | Kontextfenster | Eingabemodalitäten | Ausgabemodalitäten | Lizenz |
|---|
| Alibaba | 1.5-72b | 2024-02-25 | 32K | text | text | Qwen License |
Benchmark-Leistung
| Benchmark | Ergebnis | Einheit | Notizen |
|---|
| MMLU (Massive Multitask Language Understanding) | 81.3 | % | 5-shot |
| HumanEval | 65.2 | pass@1 | — |
| GSM8K (Grade School Math 8K) | 80.9 | % | 0-shot CoT |
| MATH | 37.6 | % | 0-shot CoT |
| BBH (BIG-Bench Hard) | 76.5 | % | 3-shot CoT |
| GPQA | 39.8 | % | 0-shot |
| IFEval | 76.4 | % | prompt_strict |
| ARC | 92.6 | % | challenge |
| MUSR | 50.8 | % | 0-shot |
| WinoGrande | 83.1 | % | 0-shot |
Preise
| Eingabe | Ausgabe | Cache-Lesen | Cache-Schreiben |
|---|
| — | — | — | — |
pro Million Token
Stärken
- MMLU score 81.3, strong knowledge reasoning.
Schwächen
Anwendungsfälle
Referenzen