The AI Loadout
←Back to models
Checked Active: 2026-10-27
Nanbeige LLM Lab

Nanbeige4.3-33B (local)

A strong open coding and office-task agent model for a 24 GB GPU or a Mac with 32 GB, if you are comfortable running llama.cpp or LM Studio.

Evaluation Analysis: Weights posted on Hugging Face on Oct 24, 2026 by Nanbeige LLM Lab. A 33B model, text only, with thinking on by default and tool calling (XML tool-call format recommended). Nanbeige's own table, run at highest thinking effort, puts it ahead of Qwen3.8-27B, Muse Glimmer 30B and Gemma 4 31B on most agent benchmarks, for example 80.9 on Terminal-Bench 2.1 and 65.5 on DeepSWE, but behind Qwen3.8-27B on SWE-Bench Pro (60.3 vs 61.7) and behind Gemma 4 31B on GPQA Diamond. Treat these as vendor numbers. Turn preserve_thinking on for multi-turn tool use and coding agents. The lab says smaller models distilled from it, such as a 3B, are coming. Not in the Ollama library. Community GGUFs (bartowski) need llama.cpp b11845 or newer and also load in LM Studio.
Context Window512K tokens (524,288)
Input Price / 1MFree (local)
Output Price / 1MFree (local)
Release Date2026-10-24

Run it locally

llama-server -hf bartowski/Nanbeige_Nanbeige4.3-33B-GGUF:Q4_K_M --jinja
Ollama tag
Not in the Ollama library
LM Studio name
Search Nanbeige4.3-33B GGUF (bartowski build)
Download size by quantization
Q4_K_M 20.8 GB; Q5_K_M 25.1 GB; Q6_K 29.0 GB; Q8_0 35.2 GB; IQ3_M 17.8 GB
License
Apache 2.0

Memory needed is roughly the download size plus 2 to 4 GB for context. See how to run AI locally.

Specs and prices checked 2026-10-27 against the official page.

Used in these setups

Related skills and personas

Agent skills and bot personas that pair well with Nanbeige4.3-33B (local).

Browse full directories:All Skills (>35)·All Personas (>15)

Want tested skills and personas installed in one command? See the AI Loadout packs, from $9.