Simon Willison's Weblog

Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding

8.5内容质量
Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding

TL;DR · AI 摘要

Ornith-1.0是DeepReinforce发布的开源大模型，基于Gemma和Qwen，支持多种参数规模，在编码任务中表现优异。

核心要点

  • Ornith-1.0提供9B到397B参数版本，兼容Apache 2.0许可证
  • 模型在编码基准测试中达到同类开源模型的最先进水平
  • 实测显示其能以103 tokens/秒速度生成图像

结构提纲

按章节快速跳转。

  1. §模型发布背景

    介绍Ornith-1.0作为DeepReinforce首个开源模型的背景信息

  2. ·技术架构

    模型基于Gemma 4和Qwen 3.5预训练，提供9B到397B参数版本

  3. ›许可证兼容性

    确认Gemma 4和Qwen 3.5的Apache 2.0许可证兼容性

  4. ·实测表现

    展示模型在Datasette代码分析和图像生成中的实际应用案例

  5. ›性能指标

    记录模型以103 tokens/秒速度生成pelican图像的测试数据

  6. ·公司背景

    提及DeepReinforce的CUDA-L1论文及技术积累

思维导图

用一张图看清主题之间的关系。

查看大纲文本（无障碍 / 无 JS 友好）
  • Ornith-1.0模型
    • 技术特性
      • 多参数版本
      • Apache 2.0兼容
    • 应用案例
      • 代码分析
      • 图像生成
    • 性能指标
      • 103 tokens/秒

金句 / Highlights

值得收藏与分享的关键句。

#LLM#开源模型#编码基准#MIT许可证#DeepReinforce
打开原文

Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding

Simon Willison’s Weblog

Subscribe

#smallhead

Sponsored by:

Depot — AI agents write code in seconds. CI shouldn't make them wait minutes.

Try Depot CI

29th June 2026 - Link Blog

Ornith-1.0: Self-Scaffolding LLMs for Agentic Coding . This is an interesting new open weights (MIT licensed) model, the first model release from DeepReinforce.

[...] with variants including 9B Dense, 31B Dense, 35B MoE, and 397B MoE. Built on top of pretrained Gemma 4 and Qwen 3.5, it achieves state-of-the-art performance among open-source models of comparable size on coding benchmarks.

As far as I can tell the licenses of those underlying models is compatible with being used in this way - Gemma 4 is Apache 2.0 licensed (and not bound by the janky additional Gemma Terms of Use that afflicted the previous Gemma models) and Qwen 3.5 is Apache 2.0 licensed as well.

I've been running the model using LM Studio and the ornith-1.0-35b-Q4_K_M.gguf (20GB) GGUF, hooked up to Pi . Initial impressions are very good - it seems to be able to run the agent harness over many tool calls in a proficient way.

Here's a terminal session where I asked it to "find the code that decodes the actor cookie" and then "find the code that opens the insert dialog when thebutton is clicked" against a Datasette checkout, which it handled with ease.

I also had it draw this pelican , which came out at 103 tokens/second:

It's a little bit mangled but the pelican is clearly a pelican.

I couldn't find much information about DeepReinforce themselves. The earliest paper I could find from the was CUDA-L1: Improving CUDA Optimization via Contrastive Reinforcement Learning from June 2025.

Posted

29th June 2026

at 4:17 pm

Recent articles

  • Porting the Moebius 0.2B image inpainting model to run in the browser with Claude Code - 22nd June 2026
  • sqlite-utils 4.0rc1 adds migrations and nested transactions - 21st June 2026
  • Datasette Apps: Host custom HTML applications inside Datasette - 18th June 2026

#primary

This is a link post by Simon Willison, posted on 29th June 2026 .

ai

2,090

generative-ai

1,847

local-llms

159

llms

1,815

qwen

57

pelican-riding-a-bicycle

120

gemma

16

llm-release

208

lm-studio

20

Monthly briefing

Sponsor me for $10/month and get a curated email digest of the month's most important LLM developments.

Pay me to send you less!

Sponsor & subscribe

.metabox

#secondary

#wrapper

  • Disclosures
  • Colophon
  • ©
  • 2002
  • 2003
  • 2004
  • 2005
  • 2006
  • 2007
  • 2008
  • 2009
  • 2010
  • 2011
  • 2012
  • 2013
  • 2014
  • 2015
  • 2016
  • 2017
  • 2018
  • 2019
  • 2020
  • 2021
  • 2022
  • 2023
  • 2024
  • 2025
  • 2026