ModelsOfficial10月26日 14:57

BOSS直聘旗下南北阁团队开源 Nanbeige4.3-33B：512K 上下文编程模型

BOSS Zhipin's Nanbeige Team Open-Sources Nanbeige4.3-33B, a 512K-Context Agentic Coding Model

官方一手Pandaily@[email protected] (Pandaily)

Blog/Media· 2 sources

Why we picked it

BOSS直聘母公司把自家的33B编程模型开源了，512K上下文还能当教师模型蒸馏出3B小版本，做智能体编程的可以拿来试试。

看准（Kanzhun）旗下南北阁（Nanbeige）团队在 Hugging Face 发布开源模型 Nanbeige4.3-33B。该模型主打 Agentic 编程和办公场景，上下文窗口达 512K。定位为教师模型，后续将蒸馏出 3B 版本。模型采用 Apache 2.0 许可证开源。

Original · Pandaily

BOSS Zhipin's Nanbeige Team Open-Sources Nanbeige4.3-33B, a 512K-Context Agentic Coding Model

The Nanbeige team at BOSS Zhipin parent Kanzhun released Nanbeige4.3-33B on Hugging Face, a teacher model for agentic coding and office work that will be distilled into a 3B version.

karminski-牙医 (AI工具)02:30原文

#Nanbeige #BOSS直聘 #开源模型 #编程助手 #智能体

Keep reading

Recent stories with the same tag or section

Models US vs ChinaSelected10:31

BOSS Zhipin releases Nanbeige4.3-33B open model

The simple Llama-style architecture means almost any inference engine can run it out of the box, at the cost of high memory usage and no built-in speculative decoding beyond external MTP. Agent-focused teams looking for an open teacher model to distil from should benchmark it against Qwen3.8-27B now, and watch for the promised sub-10B distilled releases. BOSS Zhipin being a recruiter entering open-weight modelling is itself notable for the China AI ecosystem.

#Nanbeige #BOSS Zhipin #open-source models #agents

Event

karminski-牙医 (AI工具)@karminski32 sourcesSource

Models BuildersOfficial06:03

Chinese contentNanbeige 发布 Nanbeige4.3-33B 开源智能体编程模型

Nanbeige 放出一个 33B 的开源编程模型，Apache 2.0 协议可商用，512K 上下文，Terminal-Bench 2.1 拿到 80.9 分，自己部署的可以试试。

#Nanbeige #Terminal-Bench #开源模型 #编程助手

官方一手marktechpost@Asif RazzaqSource

Models US vs China01:07

Chinese contentNanbeige 发布 Nanbeige4.3-33B 开源模型，称 ~30B 级别 SOTA

Nanbeige 刚放出 33B 开源模型，自称在多数基准上超过了 Qwen3.8-27B 和 Gemma4-31B，跑智能体编程任务的可以看看。

#Nanbeige #开源模型 #Qwen #编程助手

Event

TestingCatalog@testingcatalog2 sourcesSource

Models US vs China22:49

Chinese contentNanbeige 发布 Nanbeige4.3-33B 开源模型

Nanbeige 放了个 33B 开源模型，写代码和跑智能体比同尺寸的 Qwen 和 Gemma 都强，上下文还拉到 512K，权重随便商用。

#Nanbeige #Qwen #Gemma #开源模型

kimmonismus@kimmonismusSource

Models US vs China22:15

Chinese contentBOSS直聘旗下南北阁开源 Nanbeige4.3-33B 稠密模型

BOSS直聘团队悄悄放了个 33B 稠密开源模型，号称分数超过 Qwen3.8-27B，还有 512K 上下文，等你来验货。

#Nanbeige #Qwen #开源模型 #稠密模型

Gorden Sun@Gorden_SunSource

Papers BuildersSelected02:15

Nanbeige 4.2 Looped Transformer Breakdown Compares Weight Reuse

The three reuse tricks all called 'looped transformers' differ in cache behaviour and cost, and picking the wrong one can burn a training week or budget a 3B decode like a 3B file. Builders planning efficient architectures can use this to decide whether extra loops justify their compute and whether a second cache is needed. Nanbeige 4.2 offers a concrete open reference point against standard Transformers.

#Nanbeige #looped transformer #open-source models #KV cache

官方账号AlphaSignal@AlphaSignalAISource