OpenAI拥有推出 Pando，这是一个新的Git后端，现在服务于其单列波的所有开发。 OpenAI 潘多语 Git后端保留标准的Git客户端和协议。 工程师和Codex agents仍然像以前一样运行 git push 和 git 获取。 只有服务器改变了。 OpenAI团队报告中位推延性下降近80%. 潘多现在每天提供几乎一个Petabyte的Git数据。
TL；数据
- 比例：每日将近1个PB基特数据；高峰1小时内超过57个TB。
- 运行于：商品硬件；标准云VMs饱和10Gbps链接在克隆和获取上。
- 性能：按OpenAI本身的测试，数千个同时推力器没有退化。
- 最佳：中度推延性下降近80%(约5x).
- 最糟的是：不是开源；重复分支推动重传非Trunk历史(p99 ack under 3 MiB).
OpenAI潘多是什么来着？
Pando是来自OpenAI的Git兼容服务器后端，用于对单列波进行同步推拉和获取。 它以Rust写成，并在10周内达到生产。 一名工程师设计了Codex agents的核心和团队执行。 它自2026年4月起服务于生产流量。
OpenAI为何需要一个新的Git后端？
Codex 乘以并行工作。 兹Q 1Z的承诺图显示承诺在Q3 2026年对Q1 2024年的比分上落在主人身上1 378%. 它的Q4 2026估计值为+2,442%. 每一个推力现在都会触发很多CI抓取。 到2025年末，采集和推力放缓，合并失败，可靠性有时下降至90%以下。 OpenAI的结论是瓶颈是如何为吉特服务，而不是吉特本身。
潘道如何办？
Pando 放弃了大多数 Git 服务器视为要求的2个假设：
1. 存储器全范围的一致性：常规服务器订购每个 ref 更新全范围，如此无关的推力等待。 潘多使用因果一致性代替。 每个分支单独进行序列化，客户仅在推出同一回文时才有冲突。 取舍：浏览多个分支的读者可能在不同的时间点看到它们。
2. 每个客户端独有的包文件：标准服务器每获取一个新包。 潘多把每个推包都装成一个不可改变的包 “ref版本”进入可扩展元数据储存库。 推动需要 O(1) 元数据读写。 获取需要 O( log QQ) 读取，其中 QQ 是折叠版本的空白。
如何使推力保持小和可重复使用？
拥有上百万个分公司，一个完全校对的广告可以超过100 MiB. 潘多只广告树干分支作为新特色分支推车的基地。 任何有树干历史的阅读器都可以从一个存储的包中取出树枝在OpenAI的单列波中，中位包低于20KiB.
如何按照线性收费标准收费？
Pando通过 packfile-uri 返回下载链接，这是一个已有的 Git 协议功能。 它发送一个空的内装包。 一种名为Express缓存的快递服务区域包，使用一贯的散列，命中率超过99%.
以什么取代垃圾收集？
每次推动都会产生一个“L0”包。 背景工人群集成几何形更大的封面，类似于LSM-树皮紧凑。 单列波目前使用的封面最高可达L4. 封面起到跳过列表的作用，所以包选择以对数时间运行。
生产编号是什么？
以下所有数字均由OpenAI自行报告。
- 中间推延性下降了近80%。
- 每日近1个PB的Git数据，只使用少数CPU核心。
- 57个以上的TB包在一小时内完成，超过99%来自本地磁盘缓存。
- 克隆和获取饱和 10 Gbps 链接在标准云VM。
OpenAI还针对3相承诺和写-ahead-log设计进行了1000个写作模拟。 它标记这些教学模型，而不是基准测试。
潘多与其他方法有何比较？
潘道是怎样建造和部署的？
OpenAI为agents写了Git兼容套件，并增加了定型模拟测试。 Agents使用GPT-6 Astra写了Lean和TLA+的光谱来寻找bug.
Pando首先作为读写代理运行，因此流量没有中断时间。 OpenAI将自己的Git建造用于开发机和ChatGPT桌面应用。
回到Git上游会怎样？
OpenAI固定客户端处理同时，中断并恢复包下载，并添加包文件-uri跟踪。 它建造了新增平行的包索引和“显性资产”。 几种解决办法是上游的。 其余的坐在公共补丁追踪器中。
有什么限制？
- Pando不是开源；OpenAI团队表示正在探索分享技术的方法。
- 跨处读物没有单一的点点时快照。
- 重复或堆放的树枝将树干外的所有东西重新发送 p99包在3米B以下。
–每个性能号码都是自报的，还没有独立的基准测试。
关键外卖
- Pando保留了Git，但为平行的agent流量重建服务器。
- 每个分支的因果一致性消除了整个存储器的写法争议。
- 储存的、可重复使用的包取代每个客户包的生成。
- 中间推延性下降近80%；每日流量接近1PB。
——客户端修复向上游；后端暂时保持内部。
请检查access-date=中的日期值 (帮助) OpenAI工程站，OpenAI Git仓库和补丁跟踪器。 并加入我们的通讯。 等等你在电报里吗现在你也可以加入我们的电报。
需要和我们合作推广您的 GitHub Repo OR Hugging Face Page OR Product Release OR Webinar 等吗？联系我们
米哈尔·萨特(Michal Sutter)是一名数据科学专业人员，拥有帕多瓦大学的数据科学硕士学位。 在统计分析、机器学习和数据工程方面有一个坚实的基础，Michal在将复杂的数据集转化为可操作的见解方面表现突出。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
OpenAI has introduced Pando, a new Git backend that now serves all development in its monorepo. The OpenAI Pando Git backend keeps the standard Git client and protocol. Engineers and Codex agents still run git push and git fetch as before. Only the server changed. OpenAI team reports median push latency fell nearly 80%. Pando now delivers nearly a petabyte of Git data per day.
TL;DR
- Scale: Nearly 1 PB of Git data per day; peak above 57 TB in 1 hour.
- Runs on: Commodity hardware; standard cloud VMs saturate 10 Gbps links on clone and fetch.
- Performance: Thousands of concurrent pushers without degradation, per OpenAI’s own testing.
- Best: Median push latency down nearly 80% (about 5x).
- Worst: Not open source; repeat branch pushes resend non-trunk history (p99 pack under 3 MiB).
What is OpenAI Pando?
Pando is a Git-compatible server backend from OpenAI that scales concurrent pushes and fetches for a monorepo. It was written in Rust and reached production in 10 weeks. One engineer designed the core and teams of Codex agents implemented it. It has served production traffic since April 2026.
Why did OpenAI need a new Git backend?
Codex multiplied parallel work. OpenAI’s commit chart shows commits landed on master up 1,378% in Q3 2026 versus Q1 2024. Its Q4 2026 estimate is +2,442%. Each push now triggers many CI fetches. By late 2025, fetches and pushes slowed, merges failed, and reliability at times fell below 90%. OpenAI concluded the bottleneck was how it served Git, not Git itself.
How does Pando work?
Pando drops 2 assumptions that most Git servers treat as requirements:
1. Repository-wide consistency: Conventional servers order every ref update repo-wide, so unrelated pushes wait. Pando uses causal consistency instead. Each branch is serialized on its own, and clients only conflict when pushing the same ref. The tradeoff: a reader viewing several branches may see them at different points in time.
2. A unique packfile per client: Standard servers build a fresh pack per fetch. Pando stores each pushed pack as an immutable blob. A “ref version” goes into a scalable metadata store. A push needs O(1) metadata reads and writes. A fetch needs O(log Δ) reads, where Δ is the ref-version gap to close.
How are pushes kept small and reusable?
With millions of branches, a full ref advertisement can exceed 100 MiB. Pando advertises only the trunk branch as the base for new feature-branch pushes. Any reader with trunk history can then fetch that branch from 1 stored pack. In OpenAI’s monorepo, the median pack was under 20 KiB.
How are fetches served at line rate?
Pando returns download links via packfile-uri, an existing Git protocol feature. It sends an empty inline pack. A delivery service called Express caches packs regionally, using consistent hashing, with hit rates above 99%.
What replaces garbage collection?
Each push creates an “L0” pack. Background workers group packs into geometrically larger covers, similar to LSM-tree compaction. The monorepo currently uses covers up to L4. Covers act as a skip list, so pack selection runs in logarithmic time.
What are the production numbers?
All figures below are self-reported by OpenAI.
- Median push latency reduced by nearly 80%.
- Nearly 1 PB of Git data per day, using only a handful of CPU cores.
- Over 57 TB of packs served in a single hour, more than 99% from local disk cache.
- Clones and fetches saturate 10 Gbps links on standard cloud VMs.
OpenAI also ran a 1,000-writer simulation against 3-phase-commit and write-ahead-log designs. It labels these teaching models, not benchmarks.
How does Pando compare with other approaches?
How was Pando built and deployed?
OpenAI wrote a Git conformance suite for agents and added deterministic simulation testing. Agents using GPT-6 Astra wrote Lean and TLA+ specs to find bugs.
Pando first ran as a read- and write-through proxy, so traffic moved without downtime. OpenAI ships its own Git build to developer machines and the ChatGPT desktop app.
What goes back to upstream Git?
OpenAI fixed client handling of concurrent, interrupted and resumed pack downloads, and added packfile-uri tracing. Its build adds parallel pack indexing and “explicit haves.” Several fixes are upstream. The rest sit in a public patch tracker.
What are the limitations?
- Pando is not open source; OpenAI team states that it is exploring ways to share the technology.
- Cross-branch reads have no single point-in-time snapshot.
- Repeat or stacked branch pushes resend everything outside trunk. The p99 pack was under 3 MiB.
- Every performance number is self-reported, with no independent benchmark yet.
Key Takeaways
- Pando keeps Git, but rebuilds the server for parallel agent traffic.
- Per-branch causal consistency removes repository-wide write contention.
- Stored, reusable packs replace per-client pack generation.
- Median push latency dropped nearly 80%; daily traffic nears 1 PB.
- Client fixes go upstream; the backend stays internal for now.
Check out the full OpenAI engineering post, the OpenAI Git repository and the patch tracker. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us
Michal Sutter is a data science professional with a Master of Science in Data Science from the University of Padova. With a solid foundation in statistical analysis, machine learning, and data engineering, Michal excels at transforming complex datasets into actionable insights.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.