垂直专项评测
模态: 文本嵌入 (Embedding)
更新频率: weekly
● 自动定时拉取
MTEB Massive Text Embedding Benchmark
文本嵌入与表征向量模型标准评测体系，覆盖多任务检索、分类、聚类与语义相似度。
💡 评测口径提示： 向量模型选型必备基准，直连 MTEB 官方后端 API。
最后核验：2026-10-26
模型最新排名快照
共收录 15 款模型。数据均保留原榜单给出的分数与分项维度。
| 名次 | 模型名称 | 统一标识 | 机构 | 核心分值 | 分项指标明细 | 备注说明 | 跨榜对比 |
|---|---|---|---|---|---|---|---|
| 1 | jcorners/ingot-8b-r3 | jcorners-ingot-8b-r3 | jcorners | 75.98 | Mean Task Score: 75.98 Total Params: 7.57B Retrieval: 70 Classification: 90.41 STS: 89.32 | 发布时间: 2026-05-23, 机构: jcorners, 参数量: 7.57B | 对比 → |
| 2 | Kingsoft-LLM/QZhou-Embedding | kingsoft-llm-qzhou-embedding | Kingsoft-LLM | 75.97 | Mean Task Score: 75.97 Total Params: 7.07B Retrieval: 67.12 Classification: 88.97 STS: 91.65 | 发布时间: 2025-08-24, 机构: Kingsoft-LLM, 参数量: 7.07B | 对比 → |
| 3 | Qwen/Qwen3-Embedding-8B | qwen-qwen3-embedding-8b | Qwen | 75.22 | Mean Task Score: 75.22 Total Params: 7.57B Retrieval: 69.44 Classification: 90.42 STS: 88.58 | 发布时间: 2025-06-05, 机构: Qwen, 参数量: 7.57B | 对比 → |
| 4 | ByteDance-Seed/Seed1.5-Embedding | bytedance-seed-seed1.5-embedding | ByteDance-Seed | 74.76 | Mean Task Score: 74.76 Total Params: N/A Retrieval: 67.45 Classification: 89.88 STS: 87.23 | 发布时间: 2025-04-25, 机构: ByteDance-Seed, 参数量: N/A | 对比 → |
| 5 | infgrad/Jasper-Token-Compression-600M | infgrad-jasper-token-compression-600m | infgrad | 74.75 | Mean Task Score: 74.75 Total Params: 0.61B Retrieval: 66.19 Classification: 90.35 STS: 88.79 | 发布时间: 2025-11-14, 机构: infgrad, 参数量: 0.61B | 对比 → |
| 6 | Qwen/Qwen3-Embedding-4B | qwen-qwen3-embedding-4b | Qwen | 74.6 | Mean Task Score: 74.6 Total Params: 4.02B Retrieval: 68.46 Classification: 89.83 STS: 88.72 | 发布时间: 2025-06-05, 机构: Qwen, 参数量: 4.02B | 对比 → |
| 7 | annamodels/LGAI-Embedding-Preview | annamodels-lgai-embedding-preview | annamodels | 74.12 | Mean Task Score: 74.12 Total Params: 7.11B Retrieval: 66.18 Classification: 89.97 STS: 86.69 | 发布时间: 2025-06-11, 机构: annamodels, 参数量: 7.11B | 对比 → |
| 8 | Bytedance/Seed1.6-embedding | bytedance-seed1.6-embedding | Bytedance | 74.07 | Mean Task Score: 74.07 Total Params: N/A Retrieval: 64.9 Classification: 92.42 STS: 86.87 | 发布时间: 2025-06-18, 机构: Bytedance, 参数量: N/A | 对比 → |
| 9 | codefuse-ai/F2LLM-4B | codefuse-ai-f2llm-4b | codefuse-ai | 73.67 | Mean Task Score: 73.67 Total Params: 4.02B Retrieval: 59.63 Classification: 91.68 STS: 84.2 | 发布时间: 2025-09-18, 机构: codefuse-ai, 参数量: 4.02B | 对比 → |
| 10 | google/gemini-embedding-001 | google-gemini-embedding-001 | 73.3 | Mean Task Score: 73.3 Total Params: N/A Retrieval: 64.35 Classification: 90.05 STS: 85.29 | 发布时间: 2025-07-14, 机构: google, 参数量: N/A | 对比 → | |
| 11 | codefuse-ai/F2LLM-v2-14B | codefuse-ai-f2llm-v2-14b | codefuse-ai | 73.08 | Mean Task Score: 73.08 Total Params: 13.99B Retrieval: 60.63 Classification: 91.68 STS: 84.55 | 发布时间: 2026-03-09, 机构: codefuse-ai, 参数量: 13.99B | 对比 → |
| 12 | codefuse-ai/F2LLM-v2-8B | codefuse-ai-f2llm-v2-8b | codefuse-ai | 72.86 | Mean Task Score: 72.86 Total Params: 7.57B Retrieval: 59.82 Classification: 91.52 STS: 84.65 | 发布时间: 2026-03-09, 机构: codefuse-ai, 参数量: 7.57B | 对比 → |
| 13 | codefuse-ai/F2LLM-v2-4B | codefuse-ai-f2llm-v2-4b | codefuse-ai | 72.41 | Mean Task Score: 72.41 Total Params: 4.02B Retrieval: 59.13 Classification: 91.68 STS: 84.22 | 发布时间: 2026-03-09, 机构: codefuse-ai, 参数量: 4.02B | 对比 → |
| 14 | ai-sage/Giga-Embeddings-instruct-10B-A1.8B-0826 | ai-sage-giga-embeddings-instruct-10b-a1.8b-0826 | ai-sage | 72.23 | Mean Task Score: 72.23 Total Params: 10.48B Retrieval: 60.03 Classification: 91.62 STS: 84.27 | 发布时间: 2026-08-19, 机构: ai-sage, 参数量: 10.48B | 对比 → |
| 15 | codefuse-ai/F2LLM-1.7B | codefuse-ai-f2llm-1.7b | codefuse-ai | 72.01 | Mean Task Score: 72.01 Total Params: 1.72B Retrieval: 57.83 Classification: 90.86 STS: 83.9 | 发布时间: 2025-09-18, 机构: codefuse-ai, 参数量: 1.72B | 对比 → |
快照采集说明： 直接实时抓取自 MTEB (Massive Text Embedding Benchmark v2) 官方 Space 后端服务，涵盖 349 个文本向量嵌入与检索模型评测。