Model Qwen3-30B-A3B-Thinking-2507-ECHO-Sokoban-GRPO

Based on Qwen3-30B-A3B-Thinking, we perform GRPO training on the Sokoban dataset using the ECHO framework. Specifically, we employ a multi-round RL training schedule capped at 2 rounds, with a maximum of 50 candidate actions per round. The detailed environment configuration is as follows:

  LargerSokoban6:
    env_type: sokoban
    max_actions_per_traj: 100
    env_instruction: "You are solving the Sokoban puzzle. You are the player and you need to push all boxes to targets. When you are right next to a box, you can push it by moving in the same direction. You cannot push a box through a wall, and you cannot pull a box. The answer should be a sequence of actions, like <answer>Right || Right || Up</answer>"
    max_tokens: 300
    env_config:
      dim_x: 6
      dim_y: 6
      num_boxes: 2
      max_steps: 300
      search_depth: 20

Tabel 1: Model performance on Sokoban task

Model	Success Rate(%)
Qwen3-4B	21.8
Qwen3-4B-Echo(GRPO)	34.0
Qwen3-30B-A3B-Thinking-2507	72.75
Qwen3-30B-A3B-Thinking-2507-Echo(GRPO)	82.80
Deepseek-R1	75.75
Qwen3-235B-A22B-Thinking-2507)	79.68
gpt-oss-120b	79.69

Quick start

from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "GradientResearch/Qwen3-30B-A3B-Thinking-2507-ECHO-Sokoban-GRPO"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

# prepare the model input
prompt = "sokoban"
messages = [
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=32768
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

# parsing thinking content
try:
    # rindex finding 151668 (</think>)
    index = len(output_ids) - output_ids[::-1].index(151668)
except ValueError:
    index = 0

thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")
content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n")

print("thinking content:", thinking_content) # no opening <think> tag
print("content:", content)

Citation

If you find our work helpful, feel free to give us a cite.

@misc{xiao2025echodecouplinginferencetraining,
      title={Echo: Decoupling Inference and Training for Large-Scale RL Alignment on Heterogeneous Swarms}, 
      author={Jie Xiao and Changyuan Fan and Qingnan Ren and Alfred Long and Yuchen Zhang and Rymon Yu and Eric Yang and Lynn Ai and Shaoduo Gan},
      year={2025},
      eprint={2508.05387},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2508.05387}, 
}