可进入对比
档案摘要 / Decision snapshot
先判断能不能用于选型。
有真实任务证据
来自模型卡
边界优先
30 秒结论
可进入选型对比,但仍以案例证据为准。
来自公开模型资料库,并已补厂商/官方/模型家族证据;当前站内暂无可核验 A 类案例。
这页有来源、有风险说明、有公开可核验 A 类案例。
数据可信度 / 证据强度
强
A 类案例
5
AA 评分
18.84
官方来源
3 个补充入口
案例来源
5 条可核验
来自公开资料与模型数据库 + Artificial Analysis + 3 个厂商/官方/模型家族证据入口;AA 不是唯一事实来源。
Evidence Distribution
案例证据分布
DeepSeek R1 (Jan) 当前关联 5 条 A 类案例;这里按任务、来源和复核时间观察证据结构。
Evidence Snapshot
证据快照覆盖
DeepSeek R1 (Jan) 的 5 条 A 类案例中,已有 4 条完成本地证据快照;1 条仍在快照队列。
可用 manifest 复核
等待落盘
需补抓
需人工复核
证据完整度
5 / 5
Feishu Bitable
3 个入口
5 条
已绑定
已标注
能力边界
用标签表达信号,不用标签替代证据。
通用/待核验
暂无数据
待核验
开放/混合信号
中
| 厂商 | DeepSeek |
|---|---|
| 厂商页 | DeepSeek |
| 发布时间 | 2025-01-20 |
| 模型 ID | deepseek-r1-jan |
| 上下文 | 暂无数据 |
| 输出 | 暂无数据 |
| 模态 | 文本 / 多模态待核验 |
| 推理 | 是 / 推理模型或 thinking 模式 |
| 价格 | 官方未披露 / 暂无数据 |
| 可用平台 | DeepSeek |
| AA 评分 | 18.84 |
| 基础模型发布时间 | Artificial Analysis 未披露 |
| 资料来源 | 公开资料、厂商信息和案例库 |
| A 类案例 | 5 |
谱系位置
它在厂商路线中的位置
此页将 DeepSeek R1 (Jan) 放入 DeepSeek 的当前 Atlas 路线中。若同厂前后代资料不足,先保留为路线追踪入口,不脑补谱系关系。
2025-01-20
DeepSeek R1 (Jan) 资料状态
来自公开资料与模型数据库 + Artificial Analysis + 3 个厂商/官方/模型家族证据入口;AA 不是唯一事实来源。
公开档案
模型状态
DeepSeek R1 (Jan) 当前标记为已有真实案例。
案例库
公开案例复核
已有 5 条可核验 A 类案例,达到完整补齐线
案例时间线
最近进入档案的 A 类案例
用采集时间呈现证据进入 Atlas 的顺序,方便复核来源新鲜度。
2026-06-27T02:15:00Z
LMSYS/SGLang Team Deploys DeepSeek R1 on NVIDIA GB200 NVL72 with 3.8x/4.8x Throughput Gains
LMSYS Org / SGLang Team / infrastructure_deployment
2026-06-27T02:15:00Z
TradingAgents: 88K-Star Multi-Agent Financial Trading Framework with DeepSeek R1 Reasoning Backbone
TauricResearch / financial_analysis
2026-06-27T01:40:00Z
DeepSeek-R1-Safe: Safety-Aligned DeepSeek R1 Fine-Tuned on Huawei Ascend 910B
浙江大学网络空间安全学院 (ZJU AI Safety) + 华为 (Huawei) / model_training
2026-06-27T01:40:00Z
ChatPDF-RAG: Local PDF Question-Answering with DeepSeek R1 via Ollama
paquino11 / domain_application
2026-06-27T00:00:00Z
IntelliBar: Mac AI Assistant Leveraging DeepSeek R1 for App-Level Intelligence
IntelliBar (intellibar.app) / productivity_assistant
适合场景
- - 作为完整模型索引和厂商路线追踪入口
- - 推理/agentic workflow 候选
不适合
- - 缺少真实案例时不要包装成推荐
- - 缺字段时不要脑补价格、性能或上下文
真实案例
可核验 A 类案例精选
当前展示 5 条代表案例;排序优先看模型卡精选、证据可信度、快照状态和公开仓库/技术任务信号。案例总数仍为 5。
Representative Evidence
代表案例排序
当前第一代表案例是“LMSYS/SGLang Team Deploys DeepSeek R1 on NVIDIA GB200 NVL72 with 3.8x/4.8x Throughput Gains”。它会优先出现在本页,是因为该案例同时进入模型卡精选、具备 100/100 的证据可信度,并带有 已快照证据。
模型卡精选优先
A+ 完整链路
archive/evidence/deepseek-r1-jan-lmsys-gb200-deployment/manifest.json
Coding / repo / agent
LMSYS Org / SGLang Team 使用 DeepSeek R1 (Jan) 处理真实任务执行
LMSYS Org / SGLang Team · DeepSeek R1 (Jan)
A+ 完整链路 · 代码仓库证据
LMSYS Org / SGLang Team 公开的真实任务执行案例,来源为 官方页面,复核于 2026-06-27T02:15:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。
任务:这是一个围绕真实任务执行的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:The SGLang team at LMSYS deployed and optimized DeepSeek R1 (and V3) inference on NVIDIA GB200 NVL72 hardware, implementing prefill-decode disaggregation with FP8 attention, NVFP4 MoE quantization, and large-scale exper…
公开产物:公开材料提供原始证据链接和可访问产物,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Achieved 26,156 input tokens/sec and 13,386 output tokens/sec per GPU on DeepSeek R1 for 2000-token input sequences — a 3.8x prefill and 4.8x decode throughput improvement over H100 baseline. With BF16 attention and FP8…
模型作用:DeepSeek R1 (Jan) 在该案例中承担真实任务执行相关的生成、分析、编排或实现角色。 原始资料写作:DeepSeek R1's Mixture-of-Experts architecture with fine-grained expert parallelism was the central target model; the 671B-parameter MoE design enabled the massive throughput gains on GB200 NVL72 through large-scale expe…
A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。
风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Blog post describes benchmark/optimization results from a research team rather than end-user production deployment; however LMSYS is a major systems research organization operating real GPU clusters, and the results are…
原始记录:LMSYS/SGLang Team Deploys DeepSeek R1 on NVIDIA GB200 NVL72 with 3.8x/4.8x Throughput Gains
TauricResearch 使用 DeepSeek R1 (Jan) 处理研究分析和报告生成
TauricResearch · DeepSeek R1 (Jan)
A+ 完整链路 · 代码仓库证据
TauricResearch 公开的研究与报告生成案例,来源为 公开代码库,复核于 2026-06-27T02:15:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。
任务:这是一个围绕研究分析和报告生成的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:TauricResearch built TradingAgents, an open-source multi-agent LLM framework for financial trading, integrating DeepSeek R1 as a reasoning backbone for market analysis, sentiment analysis, and trading signal generation.…
公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Open-source framework with 88,000+ GitHub stars supporting DeepSeek R1 for multi-agent financial trading workflows including Research Manager, Trader, and Portfolio Manager agents. Published research paper (arXiv:2412.2…
模型作用:DeepSeek R1 (Jan) 在该案例中承担研究分析和报告生成相关的生成、分析、编排或实现角色。 原始资料写作:DeepSeek R1 provides chain-of-thought reasoning for the framework's core agents — Research Manager (financial analysis), Trader (signal generation), and Portfolio Manager (risk assessment) — leveraging its strong reason…
A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。
风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Framework supports multiple LLM providers; DeepSeek R1 is one of several options rather than the sole model. User must configure DeepSeek R1 explicitly. Community actively requests specific DeepSeek R1 versions (issue #…
原始记录:TradingAgents: 88K-Star Multi-Agent Financial Trading Framework with DeepSeek R1 Reasoning Backbone
浙江大学网络空间安全学院 (ZJU AI Safety) + 华为 (… 使用 DeepSeek R1 (Jan) 处理真实任务执行
浙江大学网络空间安全学院 (ZJU AI Safety) + 华为 (Huawei) · DeepSeek R1 (Jan)
A+ 完整链路 · 代码仓库证据
浙江大学网络空间安全学院 (ZJU AI Safety) + 华为 (Huawei) 公开的真实任务执行案例,来源为 公开代码库,复核于 2026-06-27T01:40:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。
任务:这是一个围绕真实任务执行的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Fine-tune the full DeepSeek R1 (671B) model for enhanced safety and compliance using multi-dimensional safety training corpus, safety-supervised fine-tuning with core safety reasoning pre-alignment, and safety reinforce…
公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Open-sourced safety-enhanced DeepSeek R1 model weights published on ModelScope (modelscope.cn/models/ZJUAISafety/DeepSeek-R1-Safe). Training pipeline includes safety data generation (jailbreak attack/defense reinforceme…
模型作用:DeepSeek R1 (Jan) 在该案例中承担真实任务执行相关的生成、分析、编排或实现角色。 原始资料写作:DeepSeek R1 is the base model being safety-fine-tuned. Its reasoning capabilities (chain-of-thought, self-verification) are preserved while safety behaviors are enhanced. The project demonstrates DeepSeek R1's adaptabil…
A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。
风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Fine-tuned derivative, not vanilla DeepSeek R1. Hardware requirement is very high (64× Ascend 910B). Academic research project with Huawei partnership.
原始记录:DeepSeek-R1-Safe: Safety-Aligned DeepSeek R1 Fine-Tuned on Huawei Ascend 910B
paquino11 使用 DeepSeek R1 (Jan) 处理知识检索和问答
paquino11 · DeepSeek R1 (Jan)
A+ 完整链路 · 代码仓库证据
paquino11 公开的知识库与检索问答案例,来源为 公开代码库,复核于 2026-06-27T01:40:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。
任务:这是一个围绕知识检索和问答的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:Build a fully local ChatPDF application using RAG (Retrieval-Augmented Generation) with DeepSeek R1 running via Ollama as the reasoning model. Users can upload PDF documents and interact with them through a Streamlit ch…
公开产物:公开材料提供公开代码、README 或项目配置,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:Open-source Streamlit application (83 GitHub stars) that runs DeepSeek R1 locally via Ollama for PDF question-answering. Features include multi-PDF upload, customizable retrieval (k results, similarity threshold), memor…
模型作用:DeepSeek R1 (Jan) 在该案例中承担知识检索和问答相关的生成、分析、编排或实现角色。 原始资料写作:DeepSeek R1 serves as the local LLM backbone for reasoning over retrieved document chunks. Its chain-of-thought reasoning enables multi-step document comprehension and accurate answer generation from PDF content, all ru…
A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。
风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Small project, primarily a demonstration/tutorial-style application. Requires local GPU for running DeepSeek R1 via Ollama. Lower star count (83) but has complete, working codebase.
原始记录:ChatPDF-RAG: Local PDF Question-Answering with DeepSeek R1 via Ollama
IntelliBar (intellibar.app) 使用 DeepSeek R1 (Jan) 处理知识检索和问答
IntelliBar (intellibar.app) · DeepSeek R1 (Jan)
A+ 完整链路 · 官方/客户故事
IntelliBar (intellibar.app) 公开的知识库与检索问答案例,来源为 官方页面,复核于 2026-06-27T00:00:00Z。 Model Atlas 将它标记为 A 类证据,因为它同时具备具体使用者、具体任务、公开原始证据和可访问产物。
任务:这是一个围绕知识检索和问答的真实任务,公开材料可以回溯到具体使用者和具体产物。 原始资料写作:IntelliBar is a Mac desktop assistant product that specifically highlights DeepSeek R1 as a supported advanced model for context-aware tasks such as editing emails within the Mail app, summarizing articles in the browse…
公开产物:公开材料提供原始证据链接和可访问产物,可用于核验任务结果、项目形态和模型绑定关系。 原始资料写作:A commercially available macOS desktop AI assistant that integrates DeepSeek R1 with system-level accessibility APIs to provide inline AI assistance across any Mac application, including email editing, article summariza…
模型作用:DeepSeek R1 (Jan) 在该案例中承担知识检索和问答相关的生成、分析、编排或实现角色。 原始资料写作:DeepSeek R1 provides the reasoning backbone for understanding context from any Mac application and generating appropriate responses. Its reasoning capabilities enable the assistant to handle complex, multi-step user req…
A 类理由:有具体使用者、具体任务、公开原始证据和可访问产物,可公开核验。
风险边界:当前判断基于公开材料;若产物下线、仓库变更或模型参与比例仅来自作者自述,需要在引用前重新复核。 原始资料写作:Product page describes DeepSeek R1 support; exact implementation details not publicly documented in code.
原始记录:IntelliBar: Mac AI Assistant Leveraging DeepSeek R1 for App-Level Intelligence
数据缺口
只基于现有字段判断缺口;缺失项不会被猜测填充。
- - 补充方向:核验 价格 的官方/API/案例来源。
- - 补充方向:核验 上下文 的官方/API/案例来源。
- - 补充方向:核验 输出 的官方/API/案例来源。
下一步验证建议
- - 继续保留 A 类案例的原始证据、产物页和版本快照。
- - 补官方发布/API/System Card 来源;若官方未披露,继续标注“官方未披露”。
- - 把 benchmark、教程、发布文保留为背景资料,不提升为真实案例。