CoolFace
Modelpublic

AtomicChat/qwen-agentworld-35b-GGUF

sourceHugging Faceapache-2.0updated 2mo agoView on Hugging Face
1likes365downloads
Model Card

<center>

<div style="display:flex; justify-content:center; align-items:center; gap:2%; max-width:560px; margin:0 auto;"> <a href="https://atomic.chat" style="flex:0 1 auto; min-width:0;"><img src="https://huggingface.co/AtomicChat/qwen-agentworld-35b-GGUF/resolve/main/pillatomicv3.png" alt="Atomic Chat" style="width:100%; height:auto; max-width:186px;"></a> <a href="https://discord.gg/8wGSsvmg4V" style="flex:0 1 auto; min-width:0;"><img src="https://huggingface.co/AtomicChat/qwen-agentworld-35b-GGUF/resolve/main/pilldiscordv3.png" alt="Join Discord" style="width:100%; height:auto; max-width:184px;"></a> <a href="https://github.com/AtomicBot-ai/Atomic-Chat" style="flex:0 1 auto; min-width:0;"><img src="https://huggingface.co/AtomicChat/qwen-agentworld-35b-GGUF/resolve/main/pillgithubv3.png" alt="GitHub" style="width:100%; height:auto; max-width:141px;"></a> </div>

<br/>

<img src="https://huggingface.co/AtomicChat/qwen-agentworld-35b-GGUF/resolve/main/hero.png" alt="Qwen Agentworld 35B A3B" style="width:100%; max-width:100%; height:auto; margin-bottom:0.6em;"/>

<div style="display:flex; justify-content:center; gap:0.5em;"> <a href="https://huggingface.co/Qwen/Qwen-AgentWorld-35B-A3B"><strong>Base model: Qwen/Qwen-AgentWorld-35B-A3B</strong></a> </div> </center>

Qwen Agentworld 35B A3B, self-quantized to GGUF by Atomic Chat. Built straight from Qwen's original weights with a per-tensor importance matrix, so this is not a repack of somebody else's files. Runs fully offline.

Highlights

  • —34.7B parameters: the weights this repo quantizes.
  • —Context length: 262,144 tokens (256K), as published by Qwen.
  • —40 layers: Mixture-of-Experts.
  • —Modalities: the base model handles Text, Image; this repo ships text-only quants, it carries no vision projector.
  • —Full imatrix ladder: every quant is calibrated with an importance matrix.
  • —Seven Unified Domains.: A single model covers MCP (tool calling), Search, Terminal, SWE (software engineering), Android, Web, and OS, spanning both text and GUI interaction environments.
  • —Native World Model.: Environment modeling from CPT onward, not post-hoc adaptation on a general-purpose LLM.
[!NOTE] These GGUFs are self-quantized from the original weights, not a repack. The importance matrix keeps low-bit quants closer to the full-precision model.
[!IMPORTANT] Always pass --jinja so the Qwen Agentworld 35B A3B chat template is applied. Without it the model can emit malformed turns.

Model Overview

PropertyValue
Base modelQwen/Qwen-AgentWorld-35B-A3B
Parameters34.7B
Layers40
Experts256 routed (top-8)
Context length262,144 tokens (256K)
Vocabulary248,320
ModalitiesText, Image in the base model; text only in this repo, it ships no vision projector
ArchitectureMixture-of-Experts, 256 experts (top-8), 16 attention heads over 2 KV heads, Qwen3_5MoeForConditionalGeneration
This repoGGUF quants (imatrix). Quants: Q4_K_M, UD-Q4_K_XL, Q5_K_M, Q6_K, Q8_0

<img src="https://huggingface.co/AtomicChat/qwen-agentworld-35b-GGUF/resolve/main/benchmark.png" alt="Qwen Agentworld 35B A3B benchmark scores" style="width:100%; max-width:900px;"/>

Scores are Qwen's published results for the base Qwen/Qwen-AgentWorld-35B-A3B, not our own measurements. Quantization preserves the large majority of this; Q4_K_M and up stay close to full precision.

Choosing a quant

QuantSizeNotes
`Q4_K_M`21.2 GBRecommended default. Best balance of size, speed and quality.
UD-Q4_K_XL21.5 GBDynamic. Embeddings and output kept at Q8_0 for higher quality at a Q4 footprint.
Q5_K_M24.7 GBHigher quality, low loss.
Q6_K28.5 GBNear lossless, noticeably lighter than Q8_0.
Q8_036.9 GBEffectively lossless, reference quality.
[!TIP] Pick the largest file that fits your (V)RAM with room for context. Q4_K_M or UD-Q4_K_XL is the sweet spot for most setups; Q6_K or Q8_0 for maximum fidelity.

Get started

Run Qwen Agentworld 35B A3B locally with:

  • —[Atomic Chat](https://atomic.chat): the easiest path. Open the app, search AtomicChat/qwen-agentworld-35b-GGUF, pick a quant, hit Use this model.
  • —llama.cpp: llama-server -hf AtomicChat/qwen-agentworld-35b-GGUF:Q4_K_M --jinja -c 8192
  • —Ollama: ollama run hf.co/AtomicChat/qwen-agentworld-35b-GGUF:Q4_K_M
  • —LM Studio / Jan: search the repo id, download any quant.

Best practices

ParameterValue
temperature0.6
top_p0.95
top_k20

Qwen's recommended sampling configuration for Qwen/Qwen-AgentWorld-35B-A3B.

Run in llama.cpp

bash
git clone https://github.com/ggml-org/llama.cpp
cmake llama.cpp -B llama.cpp/build -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON
cmake --build llama.cpp/build --config Release -j --target llama-cli llama-server
bash
./llama.cpp/build/bin/llama-server \
    -hf AtomicChat/qwen-agentworld-35b-GGUF:Q4_K_M \
    --jinja -ngl 99 -c 8192 -fa on

How these were made

  1. 1.Download Qwen/Qwen-AgentWorld-35B-A3B (original weights).
  2. 2.Convert to f16 GGUF with llama.cpp.
  3. 3.Build an importance matrix over our calibration corpus.
  4. 4.Quantize the ladder with --imatrix.
  5. 5.UD-Q4_K_XL additionally pins the token-embedding and output tensors to Q8_0.

License

Original model by Qwen, released under the Apache 2.0 license. Full terms: Apache 2.0. Quantized by Atomic Chat.