Instructions to use eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808 with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808")
model = AutoModelForCausalLM.from_pretrained("eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808

SGLang

How to use eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808 with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808 with Docker Model Runner:
```
docker model run hf.co/eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808
```

Qwen3.5-35B-A3B GLM-5.2 multi-harness agent — step 808

This is the step-808 final checkpoint of the GLM-5.2 multi-harness agent run. It is a supervised fine-tune of Qwen/Qwen3.5-35B-A3B-Base for agentic text generation. Agent SFT over the terminal and general-agent configurations of the GLM-5.2 multi-harness dataset.

This repository contains BF16 Hugging Face safetensors exported from the finalized Megatron distributed checkpoint. It is not quantized.

Checkpoint identity

Field	Value
Hugging Face repository	`eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808`
Run name	`glm52_multi_harness`
Training iteration	808
Planned iterations / epoch	808
Position in planned epoch	100.00%
Base model	`Qwen/Qwen3.5-35B-A3B-Base`
Source checkpoint	`/KRAFTON/WORKSPACE/wbl-workspace/posttraining-2606/areal_runs/qwen35_b300/sft/glm52-multi-harness-65k-from-terminal-swe-general-step7176-tp2-ep8-cp1-gbs16-lr5e6/checkpoints/iter_0000808`
Export format	BF16 Hugging Face safetensors
Context length used for SFT	65,536 tokens

Training configuration

Hardware: one node with 8 NVIDIA B300 GPUs.
Framework: Megatron-Bridge / Megatron-Core in the NVIDIA NeMo 26.06 environment.
Parallelism: TP=2, PP=1, CP=1, EP=8, DP=4.
Micro batch size: 1 per data-parallel rank.
Global batch size: 16.
Activation recomputation: full, uniform recomputation.
MoE router fusion: enabled; shared-expert overlap, gradient-reduce overlap, and parameter-gather overlap disabled.
Precision: BF16 training/model tensors with the recipe's precision-aware optimizer state; exported weights are BF16.
Learning-rate schedule: cosine decay from a peak of 5e-6 to 5e-7.
Warmup: 158 optimizer steps for every run.
Checkpoint interval: 500 optimizer steps.
Sequence packing: Yes; 12,922 packs, 99.7234% packing efficiency.
Objective: assistant-only causal language modeling. System, user, and tool messages are context rather than prediction targets.

Training data

The materialized source mixture has 92,013 rows; 91,693 rows enter training after the 32K boundary check. The complete training representation contains 844,513,664 effective/rendered tokens.

Source dataset	Selected rows
`eewer/glm-5.2-multi-harness-agent`	92,013

The source contains 38,555 terminal-harness rows and 53,458 general-agent rows from the agentic_tool_use configuration. The combined data was deterministically shuffled with seed 20260722, then packed to 65,536 tokens. A total of 320 overlength rows were rejected. Assistant turns with empty or missing reasoning are retained and rendered as an opening <think> tag, a blank line, and a closing </think> tag before the answer content.

The tokenizer is from Qwen3.5, with a Qwen3.6 thinking-preserving chat template. The template keeps reasoning content in assistant messages and preserves tool declarations when the source supplies a validated schema.

Intended use

This checkpoint is intended for research on terminal, software-engineering, tool-use, and interactive agents. Use the included chat template and supply tool schemas expected by the target harness. It can be served with Transformers-compatible runtimes that support the Qwen3.5 MoE architecture.

This is a training checkpoint, not a polished instruction model. Compare checkpoints using held-out loss and task-level agent evaluations before selecting one for deployment.

Limitations and safety

No complete benchmark suite is claimed in this model card.
Agent trajectories can produce destructive shell commands, modify files, call tools, or expose secrets. Run the model in an isolated environment with scoped credentials.
Tool names and schemas vary across source harnesses. A caller must provide the schema appropriate for its own environment; do not assume a generated call is valid or safe.
Dataset filtering and deduplication reduce, but cannot guarantee removal of, benchmark contamination, incorrect reasoning, insecure code, or teacher-model artifacts.
The model can hallucinate successful tool execution and should not be trusted without checking actual environment observations.
This checkpoint inherits the base model's limitations and license.

Reproducibility notes

Both learning-rate variants use the same TP2/EP8 65K execution topology, optimizer family, 158-step warmup, shuffled packed dataset, and checkpoint cadence. They differ only in peak/minimum learning rate.

Downloads last month: 16

Safetensors

Model size

35B params

Tensor type

BF16

Model tree for eewer/qwen3.5-35b-a3b-glm52-multi-harness-65k-lr5e6-step808

Base model

Qwen/Qwen3.5-35B-A3B-Base

Finetuned

(77)

this model