---
title: Qwen-AgentWorld-35B-A3B-bf16
canonical_url: "https://www.modelscope.cn/models/mlx-community/Qwen-AgentWorld-35B-A3B-bf16"
md_url: "https://www.modelscope.cn/models/mlx-community/Qwen-AgentWorld-35B-A3B-bf16.md"
repository: mlx-community/Qwen-AgentWorld-35B-A3B-bf16
last_updated: 2026-06-28
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3_5_moe
architectures:
  - Qwen3_5MoeForConditionalGeneration
base_model:
  - Qwen/Qwen-AgentWorld-35B-A3B
base_model_relation: finetune
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - Pytorch
downloads: 38
stars: 0
tags:
  - mlx
  - qwen
  - world-model
  - agent
  - environment-simulation
---

# Qwen-AgentWorld-35B-A3B-bf16

> Qwen-AgentWorld-35B-A3B-bf16 - mlx-community 在 ModelScope 开源的模型。Qwen-AgentWorld-35B-A3B-bf16

mlx-community/Qwen-AgentWorld-35B-A3B-bf16 是 ModelScope 魔搭社区上的text-generation模型，采用 apache-2.0 许可，基于 Qwen/Qwen-AgentWorld-35B-A3B 构建。

- **Repository**: mlx-community/Qwen-AgentWorld-35B-A3B-bf16
- **License**: apache-2.0
- **Tasks**: text-generation
- **Base model**: Qwen/Qwen-AgentWorld-35B-A3B
- **Tags**: mlx, qwen, world-model, agent, environment-simulation
- **Downloads**: 38
- **Stars**: 0
- **Last updated**: 2026-06-28

Source: https://www.modelscope.cn/models/mlx-community/Qwen-AgentWorld-35B-A3B-bf16

---

# Qwen-AgentWorld-35B-A3B-bf16

Full-precision (bfloat16) MLX build of
[Qwen/Qwen-AgentWorld-35B-A3B](https://huggingface.co/Qwen/Qwen-AgentWorld-35B-A3B).
The original weights load in `mlx-lm` / oMLX as-is — this repo packages them for
Apple Silicon without any precision loss.

For Apple Silicon. Runs in `mlx-lm`, oMLX, or any MLX app. Quantized siblings:
[oQ4](https://huggingface.co/mlx-community/Qwen-AgentWorld-35B-A3B-oQ4)
(≈4.6 bpw, ≈19 GB) and
[oQ3.5](https://huggingface.co/mlx-community/Qwen-AgentWorld-35B-A3B-oQ3.5)
(≈3.5 bpw, ≈16 GB).

## Notes

- **Text-only.** The checkpoint declares a `vision_config` and MTP heads in
  `config.json`, but ships **no** vision or `mtp.*` weights (693 tensors, 0 vision,
  0 MTP) — both are vestigial skeleton inherited from the Qwen3.5 base, and
  `language_model_only: true` makes it a pure LM. Nothing multimodal is present to use.
- **≈65 GB** on disk; peak memory ≈69 GB generating. Fits in 128 GB unified memory;
  use a quantized sibling on smaller machines.

## Performance

Measured with [oMLX](https://github.com/jundot/omlx) (Auto engine) on a
Macbook Pro M5 Max 128GB 40 GPU. Single request:

| Context (pp/tg) | TTFT | decode | prefill | peak mem |
|---|---|---|---|---|
| 1024 / 128  | 644 ms | 77 tok/s | 1591 tok/s | 65.6 GB |
| 4096 / 128  | 1.68 s | 76 tok/s | 2434 tok/s | 66.4 GB |
| 8192 / 128  | 2.39 s | 75 tok/s | 3428 tok/s | 66.7 GB |
| 32768 / 128 | 12.0 s | 67 tok/s | 2730 tok/s | 68.7 GB |

For ≈1.8× the single-request decode throughput at ≈⅓ the memory, see the
[oQ4 sibling](https://huggingface.co/mlx-community/Qwen-AgentWorld-35B-A3B-oQ4).

## Usage

```bash
mlx_lm.generate --model mlx-community/Qwen-AgentWorld-35B-A3B-bf16 \
  --system-prompt "You are a language world model simulating a Linux terminal. Given the user's command, predict the terminal output." \
  --prompt $'Action: execute_bash\nCommand: ls -la /home/user/project/' \
  --max-tokens 512 --temp 0.6
```

The model uses thinking mode (`<think>...</think>`) by default. Recommended
sampling: `temperature=0.6`, `top_p=0.95`, `top_k=20`. See the
[base model card](https://huggingface.co/Qwen/Qwen-AgentWorld-35B-A3B) for the
seven agent domains and domain-specific system prompts.