whatmodel.sh / Qwen-AgentWorld 35B-A3B

Qwen-AgentWorld 35B-A3B

+ contribute
Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Qwen-AgentWorld 35B-A3B is an open-weights language model from Alibaba, released June 2026. Its published size is 34.66B parameters (3B active per token) and a 262k-token context window. Licence: apache-2.0.

Catalogue identity notes

Qwen/Qwen-AgentWorld-35B-A3B @60d2b04 (created 2026-06-22, licence apache-2.0). Qwen's language world model (June 2026; Qwen3.5-35B-A3B-Base, 35B total / 3B activated) trained to simulate agentic environments. Rows: its own AgentWorldBench table (new benchmark proposal), with 11 catalogued comparators. Held at the 2026-10-10 merge because its only result is Qwen's own AgentWorldBench (a test built for this model); imported under the 2026-10-11 owner decision.

Model information sources

Release date ↗ · Weights licence ↗

HF safetensors metadata 34,660,610,688 parameters; 3B active per the maker card.

72.5±4.3#62 of 949 LLMsfrom 1 benchmark result across 1 source* thinly tested, treat the rank as provisional

$ how it compares

The LLMs ranked either side of Qwen-AgentWorld 35B-A3B, out of 1132 with a score, then more from Alibaba.

#modelscorewhereRAMreleased
70LongCat 2.0 (1.6T-A48B) Meituan73±2api $1.30/M +local~2155GB2026-06
71Gemini 3 Pro (Preview) Google73±1comparison only2025-11
72Qwen-AgentWorld 35B-A3B Alibaba73±4local~22GB2026-06
73GPT-5 OpenAI72±1api $3.44/M2025-08
74Qwen3.6 Max Preview Alibaba72±1api $2.92/M2026-04
more from Alibaba
18Qwen3.8 Max (2.4T-A95B) Alibaba78±1api $3.00/M +local~380GB2026-08
34Qwen3.8 Max (0902) Alibaba76±1api $3.00/M2026-09
37Qwen3.8-Flash-Next Alibaba75±1local~59GB2026-08

see all LLMs · find the ones that fit your machine

$ what it's good at

other tests1 not in a task · 92% of its score
testresultvs the others in its rundifficultycountswhy
AgentWorldBench (open-ended, overall) lab56.4% ✎▄▄··▄▄·▄█▄██ beats 8/11hard · avg 25%▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 92%blunt · at its level

$ where to run it

llama.cpp · recommended

  1. Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
    macOS / Linux
    curl -LsSf https://llama.app/install.sh | sh
    Windows
    irm https://llama.app/install.ps1 | iex

    other ways to install ↗

  2. Run Qwen-AgentWorld 35B-A3B · needs ~22.32 GB · from huggingface.co/unsloth
    chat in the terminal
    llama cli -hf unsloth/Qwen-AgentWorld-35B-A3B-GGUF:UD-Q4_K_XL
  3. Or start a local server and open the address it prints (http://localhost:9931)
    llama serve -hf unsloth/Qwen-AgentWorld-35B-A3B-GGUF:UD-Q4_K_XL

LM Studio

version

  1. Install LM Studio · Open the app once before using the lms command.

    download it (macOS / Windows / Linux) ↗

  2. Run Qwen-AgentWorld 35B-A3B · needs ~22.32 GB · from huggingface.co/unsloth

    open in LM Studio ↗

  3. Or download it in the terminal
    lms get https://huggingface.co/unsloth/Qwen-AgentWorld-35B-A3B-GGUF@UD-Q4_K_XL -y
  4. Then run lms ls to see its name, and lms chat followed by that name

Ollama

  1. Install Ollama

    download it (macOS / Windows) ↗

    Linux
    curl -fsSL https://ollama.com/install.sh | sh
  2. Run Qwen-AgentWorld 35B-A3B · needs ~22.32 GB · from huggingface.co/unsloth
    chat in the terminal
    ollama run hf.co/unsloth/Qwen-AgentWorld-35B-A3B-GGUF:UD-Q4_K_XL

MLX (Mac)

version

  1. Install MLX (Mac) · Apple silicon Macs only, text only.
    with uv
    uv tool install mlx-lm
    or
    pip install mlx-lm
  2. Run Qwen-AgentWorld 35B-A3B · needs ~19.5 GB, Mac only · from huggingface.co/erobey
    chat in the terminal
    mlx_lm.chat --model erobey/Qwen-AgentWorld-35B-A3B-mlx-4bit
  3. Or start an OpenAI-compatible server at http://localhost:8080/v1
    mlx_lm.server --model erobey/Qwen-AgentWorld-35B-A3B-mlx-4bit

Other ways to run it

  1. Run it as published · needs ~69.3 GB, NVIDIA only · from huggingface.co/Qwen (official)
    vllm serve Qwen/Qwen-AgentWorld-35B-A3B

Official weights and documentation: huggingface.co/Qwen ↗

$ how the score was worked out · Qwen/Qwen-AgentWorld-35B-A3B… 92%

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

sourceresultsshare of the scorelicence
Qwen/Qwen-AgentWorld-35B-A3B model card · lab-reported1▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇▇ 92%Apache-2.0
prior · pull toward the middle▇▇ 8%
weighted average= 72.5

$ improve this page

Update model informationAdd a place to run itAdd a test resultPropose a benchmarkCorrect a benchmark scoreEdit data on GitHub

Use a short GitHub form or edit the data through a pull request. Every number needs a source; here's which ones we can use.

data updated · sources and licences · json