开放权重模型比较
从质量、性能、推理速度、上下文窗口、参数量和许可证详情等关键指标比较和分析开放权重 AI 模型。
如果模型权重可供下载,我们便将其视为开放权重模型。这样用户就可以在自己的基础设施上自行托管,并通过微调等方式定制模型。
如需了解我们方法论的更多详情,请参阅常见问题。
开放性
Artificial Analysis Openness Index:得分
开放权重模型进展
开放权重与专有模型:Artificial Analysis Progress Index 变化趋势
开放权重:各实验室 Artificial Analysis Progress Index 随时间的变化
开放权重:各规模 Artificial Analysis Progress Index 随时间的变化
智能
Artificial Analysis Intelligence Index
Evaluations
Agentic knowledge work, Elo
Knowledge
1 - hallucination rate
Agentic SaaS workflows
Agentic scientific research workflows in a terminal
Agentic software engineering in open-source repositories
Reasoning & knowledge
Professional document reasoning, All-pass
Agentic coding & terminal use
Physics reasoning
Quantitative analysis on spreadsheets & documents
Legal agentic work, Hallucination-Gated All-Pass Rate
Agentic business operations
Agentic real-world work tasks, Elo
Agentic interactive 3D experiences, Elo
Product and brand videos designed in code, Elo
Kubernetes incident root-cause analysis
Medical long context reasoning
Visual reasoning
Rich, varied, well-crafted language
Well-structured, scannable, answer-first
Truthful, acknowledges uncertainty, non-sycophantic
Use of filler and generic language
规模
按模型规模划分的 Intelligence Index
模型规模:总参数量与活跃参数量
Intelligence Index 与活跃参数
Intelligence Index 与总参数量
上下文窗口
上下文窗口
更多详情
权重 | 服务商基准测试 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Kimi K3.1 Pro | 46 | 2.8T 推理时启用 104B 个参数 | 1M | $1.54 | 38.7 | 31 | +15 | |||
DeepSeek V4.1 Flash 1017 (Max) | 33 | 552B 推理时启用 16B 个参数 | 1M | $0.18 | 13.1 | 177 | +23 | |||
GLM-5.3 (Max) | 32 | 753B 推理时启用 40B 个参数 | 1M | $0.90 | 22.2 | 89 | +26 | |||
Inkling 2 Small | 31 | 594B 推理时启用 17B 个参数 | 1M | $0.18 | 21.4 | 117 | ||||
MiMo-V2.6-Pro | 30 | 1T 推理时启用 42B 个参数 | 1M | $0.18 | 35.8 | 56 | ||||
Ling 3.1 Flash 1019 | 30 | 560B 推理时启用 25B 个参数 | 1M | $0.19 | - | - | ||||
Mistral Large 4 (High) | 26 | 1.1T 推理时启用 52B 个参数 | 1M | $0.40 | 18.4 | 74 | ||||
Qwen3.8 27B (Xhigh) | 20 | 27B | 256k | $0.47 | 38.8 | 47 | +7 | |||
K2 Horizon 375B A23B | 18 | 375B 推理时启用 23B 个参数 | 524k | $0.00 | 4.7 | 224 | ||||
Beam 1.0 | 18 | 501B 推理时启用 23B 个参数 | 1M | $0.23 | 2.2 | 257 | +2 | |||
Nemotron 3.5 Super VL | 14 | 120B 推理时启用 12B 个参数 | 1M | $0.13 | 4.5 | 300 | +6 |