用户问的是"Ornith-1.0是什么模型?"。这是一个我不熟悉的术语,需要进行搜索。按照规划要求,我需要编制6个搜索关键词从不同角度并行搜索。让我思考一下:
- "Ornith-1.0" - 直接搜索这个名称
- "Ornith-1.0 model" - 英文搜索,可能是一个AI模型
- "Ornith-1.0 模型" - 中文搜索
- "Ornith 1.0" - 带空格的版本
- "Ornith-1.0 AI" - 可能是AI领域的新模型
- "Ornith model" - 更宽泛的搜索
或者也可以考虑:
- 可能是一种生物相关的模型?"Ornith" 词根与鸟有关(ornithology 鸟类学)
- 可能是某种技术模型
让我用6个不同的关键词并行搜索,count=20。
🧭 需求解读与搜索规划
Ornith 词根与鸟类学(ornithology)相关,但具体是哪种模型(AI模型?生物学模型?技术框架?)尚不确定,因此我设计6个不同角度的搜索关键词并行检索,确保覆盖各种可能性。搜索结果显示,Ornith-1.0 是一个近期(2026年6月)由 DeepReinforce 发布的开源大语言模型家族,专注于 agentic coding(代理式编程)。有多个模型规模和关键特性。
我现在需要深入阅读一些关键来源来获取详细信息。让我并行阅读几个最有价值的页面:
- 官方页面(deep-reinforce.com)- 了解核心概念
- Hugging Face 集合 - 了解模型规模
- GitHub 仓库 - 了解技术细节
- Simon Willison 的文章 - 技术分析
- ExplainX 的文章 - 技术解读
- 知乎文章 - 中文解读
这些来源可以交叉验证信息。
I now have comprehensive information about Ornith-1.0 from multiple sources. Let me also read one more source to get additional details about DeepReinforce and the model's background.
Let me also check a few more sources to round out the picture - the MarkTechPost article and the explainx.ai article for additional perspectives.
Actually, I already have very comprehensive information from:
- DeepReinforce official blog - core innovation details
- Simon Willison's blog - independent analysis
- GitHub repo - benchmark data and deployment info
- Hugging Face collection - model variants
Let me also read the explainx.ai article and the verdent.ai guide for more details.
Wait, let me think about what else I need. The question is "Ornith-1.0是什么模型?" which is simple - what kind of model is it. I already have enough information to answer comprehensively. Let me also look into DeepReinforce the organization behind it, and maybe the MarkTechPost article for another perspective.
Let me read a couple more articles to get a fuller picture and then compile the report.
信息已经非常丰富了!让我再补充阅读几个不同的视角来完善理解。
I now have extremely comprehensive information about Ornith-1.0 from multiple authoritative sources. Let me verify the key facts across sources:
-
What is Ornith-1.0? - Cross-verified across all sources:
- An open-source, MIT-licensed family of large language models (LLMs) specialized for agentic coding
- Released by DeepReinforce on June 25, 2026
- Built on top of Gemma 4 and Qwen 3.5 pre-trained weights
-
Model variants - Multiple sources confirm:
- 9B Dense (≈9 billion parameters)
- 31B Dense (≈31 billion parameters)
- 35B MoE (≈35 billion parameters, Mixture-of-Experts)
- 397B MoE (≈397 billion parameters, Mixture-of-Experts - flagship)
-
Core innovation (Self-Scaffolding) - Confirmed by official blog, Simon Willison, explainx.ai, and others:
- The model learns to generate its own "scaffold" (the harness/agent framework) during RL training, not just the solution
- Two-stage RL process: first optimize the scaffold, then generate solutions using that scaffold
- Rewards back-propagated to both stages
-
Benchmark performance - Detailed data from GitHub repo and official blog:
- 397B model: Terminal-Bench 2.1: 77.5, SWE-Bench Verified: 82.4 (beats Claude Opus 4.7 in some benchmarks)
- 35B model: Terminal-Bench 2.1: 64.2 (beats Qwen 3.5-397B's 53.5)
- 9B model: Terminal-Bench 2.1: 43.1, SWE-Bench Verified: 69.4
-
Anti-reward hacking mechanisms - Three layers: fixed trust boundary, deterministic monitor, frozen LLM judge
-
Deployment - Available via vLLM, SGLang, Ollama, LM Studio, llama.cpp; MIT license, no restrictions
-
Organization behind it - DeepReinforce (AI research collective/team, first model release)
Now let me check if there are any discrepancies across sources:
- The release date: Most sources say June 25, 2026; some say June 26. Both are consistent within a narrow window.
- The model sizes: All sources agree on 9B, 31B, 35B MoE, 397B MoE
- The base models: Most say Qwen 3.5 and Gemma 4; the official blog explicitly states "built on pre-trained Gemma 4 and Qwen 3.5"
- The GitHub says 31B is built on Gemma, the others on Qwen 3.5
I think I have very comprehensive information now. Let me check if there are any remaining questions or issues:
Current status:
- ✅ Key facts verified across 2-3+ independent sources
- ✅ No unresolved contradictions found
- ✅ No next steps needed for search
- ✅ Information is sufficient to answer the question
I can now proceed to start-report to write the final report. Let me compile all the information.
Wait, let me double-check the key remaining questions:
- "仍待验证问题" - None. All key facts about Ornith-1.0 are well established.
- "下一步检索方向" - None. Information is fully sufficient.
- All plan tasks executed.
I'm ready to call start-report.
信息已经非常充实,多源交叉验证完成。现在进入最终报告撰写阶段。
Now I need to write a comprehensive search brief/report about Ornith-1.0 based on all the information I've gathered. Let me organize all the verified facts from multiple sources and create a well-structured report.
Let me compile the key information:
- An open-source, MIT-licensed family of large language models (LLMs) specialized for agentic coding
- Released by DeepReinforce on June 25, 2026
- Built on pre-trained Gemma 4 and Qwen 3.5 base models
- Named after the Greek word for "bird" (寓意像鸟儿筑巢一样自主构建框架)
- 9B Dense (≈9 billion parameters) - edge/single GPU
- 31B Dense (≈31 billion parameters) - based on Gemma
- 35B MoE (≈35 billion parameters, Mixture-of-Experts) - practical balance
- 397B MoE (≈397 billion parameters, MoE) - flagship
- The model learns to generate its own agent scaffold (harness) during RL training, not just the solution
- Two-stage RL process: first optimize the scaffold, then generate solutions using that scaffold
- Rewards back-propagated to both stages
- Anti-reward hacking: three-layer defense mechanism
- 397B: Terminal-Bench 2.1: 77.5, SWE-Bench Verified: 82.4
- 35B: Terminal-Bench 2.1: 64.2 (beats Qwen 3.5-397B's 53.5)
- 9B: Terminal-Bench 2.1: 43.1, SWE-Bench Verified: 69.4
- vLLM, SGLang, Ollama, LM Studio, llama.cpp
- MIT license, no geographical restrictions
- 262K context window
- All benchmark results are self-reported by DeepReinforce, not independently verified
- Performance depends heavily on test framework and settings
- Ornith is weights only, not a ready-to-use product
Let me write a comprehensive report now.
I need to be careful about citations. Let me map the sources:
- 74 Simon Willison's blog
- 75 Hugging Face collection
- 76 DeepReinforce official blog
- 77 知乎 (access denied)
- 78 GitHub repository
- 79 MarkTechPost
- 80 Verdent.ai guide
- 81 explainx.ai blog
Let me write the report now.
🦅 Ornith-1.0:DeepReinforce 推出的自改进开源 Agentic Coding 大模型家族
🎯 核心发现速览
🧩 模型全景:参数规格与定位
| 版本 | 架构 | 基座模型 | 最佳部署场景 |
|---|---|---|---|
| Ornith-1.0-9B | Dense(约 90 亿参数) | Qwen 3.5 | 单 GPU 部署/边缘设备 |
| Ornith-1.0-31B | Dense(约 310 亿参数) | Gemma 4 | 多 GPU 部署 |
| Ornith-1.0-35B | MoE(约 350 亿参数,每 token 激活约 30 亿) | Qwen 3.5 | 性能与效率的平衡点 |
| Ornith-1.0-397B | MoE(约 3970 亿参数,旗舰版) | Qwen 3.5 | 复杂长周期任务最高精度 |
💡 核心创新:自改进训练框架(Self-Scaffolding)
🏗️ 传统方案的局限
🔄 Ornith 的两阶段 RL 训练
- 框架阶段(Scaffold Stage) :基于当前任务和此前使用的框架,模型首先提出优化后的 Agent 框架(包括内存管理、重试策略、工具编排等)
- 解决方案阶段(Solution Stage) :基于优化后的框架和任务描述,模型生成具体的代码解决方案
🛡️ 三层奖励欺骗防御机制
- 固定信任边界 :环境、工具接口和测试隔离机制不可修改,模型仅能优化内部策略逻辑
- 确定性监控器 :基于规则的监控器标记任何读取未授权路径、修改验证脚本或调用未授权工具的行为,违规轨迹获零奖励并从优势计算中排除
- 冻结 LLM 评判器 :作为验证器之上的否决机制,拦截意图层面的作弊行为
⏳ 异步强化学习(Pipeline-RL)
📊 基准测试表现
🏆 旗舰版 Ornith-1.0-397B 的关键对比
| 基准测试 | Ornith-1.0-397B | Qwen 3.5-397B | Claude Opus 4.7 | Claude Opus 4.8 | DeepSeek-V4-Pro |
|---|---|---|---|---|---|
| Terminal-Bench 2.1 (Terminus-2) | 77.5 | 53.5 | 70.3 | 85.0 | 67.9 |
| SWE-Bench Verified | 82.4 | 76.4 | 80.8 | 87.6 | 80.6 |
| SWE-Bench Pro | 62.2 | 51.6 | 64.3 | 69.2 | 55.4 |
| NL2Repo | 48.2 | 36.8 | — | 69.7 | — |
| ClawEval (平均) | 77.1 | 70.7 | 78.2 | — | 75.8 |
⚡ 中小尺寸的"以小博大"
- Ornith-1.0-35B 在 Terminal-Bench 2.1 上得分 64.2 , 超过 Qwen 3.5-397B 的 53.5 ,展示了自框架训练对 Agent 能力的显著压缩效果767881。
- Ornith-1.0-9B 在 SWE-Bench Verified 上得分 69.4 ,基本与 Gemma 4-31B(52.0)拉开显著差距,证明 Agent 编程能力可被高效压缩到可边缘部署的模型尺寸中7678。
🚀 部署与使用
支持的运行环境
- vLLM (推荐,≥ 0.19.1)
- SGLang (≥ 0.5.9)
- Ollama (GGUF 量化版本)
- LM Studio (本地桌面运行)
- llama.cpp (GGUF 量化版本)
- Hugging Face Transformers (≥ 5.8.1)
推理模型特性
🏢 发布方:DeepReinforce
deepreinforce-ai 组织发布模型75,并在 Twitter/X 上以 @ornith_ 账号运营81。⚖️ 局限性声明与审慎提示
- 基准测试未经第三方独立验证 :所有测试结果均由 DeepReinforce 自行发布,评分高度依赖测试框架与参数设置(温度、上下文窗口、采样设置等),不同测试框架下同一模型的得分可能存在显著差异。引用时应标注为"DeepReinforce 报告的……"并注明测试条件8081。
- 仅提供模型权重 :Ornith-1.0 不是开箱即用的托管编码助手,用户需自行从 Hugging Face 下载权重,通过 vLLM/Ollama/LM Studio 等工具部署并集成到 Agent 框架中80。
- 尚在早期阶段 :作为 DeepReinforce 的首个模型发布,其社区生态、第三方工具支持、文档完善度仍在成长中74。
- 部分基准仍存差距 :在 SWE-Bench Pro 上,Ornith-1.0-397B(62.2)低于 Claude Opus 4.7(64.3);Claude Opus 4.8 和 GLM-5.2-744B 在多项指标上仍保持领先767881。
- 可用性需自行确认 :部分 checkpoint 的实际发布状态和支持的运行时版本需以官方 Hugging Face 仓库为准80。
🌟 总结与定位
在 2026 年的 Agentic Coding 赛道中,Ornith-1.0 的差异化定位是: