LocalClaw/LLM/Nanbeige4.2 3B
Open-weight local LLM

Nanbeige4.2 3B

Catalogue summary: Nanbeige compact Apache 2.0 agentic model with 256K context, strong official code-agent and office-agent claims, and practical Q4_K_M GGUF/Ollama paths through Nanbeige-compatible llama.cpp runtimes.

Repository editorial metadata; verify comparative claims in the linked upstream material.

8 GB catalogue minimum8 GB RAMQ4_K_MCoding assistant
Choose an app

This model needs a special runtime. Unsupported apps are clearly marked.

Compare models
LM StudioNot available for this model
UnslothNot available for this model
OllamaNot available for this model
Open on Hugging FaceFiles, licence and available downloads
llama.cppNot available for this model
Use with LocalClawOptional workspace after the model is installed
This model needs its official runtimeNanbeige llama.cpp / Ollama fork
Parameters
4B (3B non-embedding)
Minimum RAM
8 GB
Model size
2.2 GB
Quantization
Q4_K_M

Can Nanbeige4.2 3B run locally?

Nanbeige4.2 3B has a catalogue minimum of 8 GB RAM with Q4_K_M. Actual memory use and speed vary by context length, runtime, backend and system headroom.

Use the official Nanbeige llama.cpp / Ollama fork setup. The current low-bit files are not a stock LM Studio install.

Model sourcePublic GGUF repository (verified 2026-09-07): owao/Nanbeige4.2-3B-GGUFUpstream releasePaper or model cardLicence text
chatcodereasoningagentlightlong-contexttool-calling

Install path

01
Check RAM fitMinimum 8 GB RAM. Start with the Q4_K_M quant.
02
Load the modelFollow the official Nanbeige llama.cpp / Ollama fork instructions. Stock LM Studio support is not confirmed.
03
Control locallyUse LocalClaw to manage models, agents, chat, channels and scheduled OpenClaw work.

Catalogue record

  • Family: nanbeige
  • Parameters: 4B (3B non-embedding)
  • Recommended quantization: Q4_K_M
  • Catalogue minimum RAM: 8 GB
  • Catalogue model size: 2.2 GB
  • Tags: chat, code, reasoning, agent, light, long-context, tool-calling

Practical limits

  • Catalogue RAM is a minimum estimate, not a guarantee for every context length or runtime.
  • Speed and memory use vary by quantization, backend, context length and system headroom.
  • Verify architecture, licence and usage restrictions in the linked upstream material before deployment.

Catalogue tags

  • chat
  • code
  • reasoning
  • agent
  • light
  • long-context

Capability profile

Repository catalogue ratings used by LocalClaw's editorial rubric. They are not a standardized third-party benchmark.

speed
9
quality
7
coding
8
reasoning
8

Technical notes

Developer
Nanbeige Team
License
Apache 2.0
Context window
262,144 tokens
Architecture
Looped Transformer causal language model with about 4B total parameters and 3B non-embedding parameters. The architecture reuses transformer layers and adds Nanbeige-specific runtime code.

This model fits these next steps

Hardware fit is based on LocalClaw's RAM tier, model size and quantization metadata. Always leave memory headroom for your OS and runtime.

Entry laptop fitMacBook Air 8GBMore headroomMac mini M4 16GBAll compatible picks8GB RAM guide

Related catalogue entries

Linked mechanically by family, shared tags and nearby RAM tier; this is not a quality ranking.

MiniCPM5 2B 2BAgents-A1 4B 4BSpark-X2.5-1.7B 1.7BGranite 4.2 (3B) 3BK2-Horizon-0.9B 0.9BLFM2.5-2.6B 2.7BLing-3.0-tiny 7.9B (1.3B active, MoE)Granite 4.1 (3B) 3B

Where to go next