
Base model	Qwen/Qwen3-Coder-Next
Format	BF16
Total params	57B
Active / token	—
Experts / layer	359
Layers	48
Hidden size	2048
Context	262,144
On-disk size	113 GB

Which variant should I pick?

Variant	Format	Link
`Qwen3-Coder-57B` (this)	BF16	link
`Qwen3-Coder-64B`	BF16	link

30% expert-pruned version of Qwen/Qwen3-Coder-Next using Cerebras REAP (Router-weighted Expert Activation Pruning).

	Original	This Model
Total params	~80B	56.56B
Experts	512	359
Active params/tok	~4.2B	~4.2B
Experts/tok	10	10
Format	BF16	BF16
Disk size	~149 GB	~113 GB

REAP removes 30% of MoE experts (153 of 512) while preserving the model's routing behavior and output quality. The active parameter count per token is unchanged since the router still selects 10 experts per token from the remaining pool. This yields a ~24% reduction in total disk/memory footprint at the cost of moderate quality degradation, primarily in math tasks.

Method

REAP (ICLR 2026) prunes Mixture-of-Experts models by scoring expert importance using:

Router gate values -- how often and how strongly the router selects each expert
Expert activation norms -- magnitude of each expert's output contribution
Frequency-weighted saliency -- combining routing frequency with activation importance
Router logit renormalization -- maintains output distribution after expert removal
Layerwise application -- independent per-layer pruning decisions for stability

Calibration Dataset

22,000 samples (no-refusal subset: 21,000), packed to 16,384 token sequences:

Category	Samples	Source
Coding (general)	4,096	`theblackcat102/evol-codealpaca-v1`
Reasoning (code)	~2,680	`open-r1/Mixture-of-Thoughts[code]`
Reasoning (math)	~2,778	`open-r1/Mixture-of-Thoughts[math]`
Reasoning (science)	~2,776	`open-r1/Mixture-of-Thoughts[science]`
Tool calling	4,096	`Salesforce/xlam-function-calling-60k`
Agentic coding	4,096	`SWE-bench/SWE-smith-trajectories`
+ extended domains	~1,478	Scientific, CUDA kernels, browser, advanced math, code correctness

Total tokens observed: ~90.5M across 6,391 packed sequences.

Pruning Configuration

Parameter	Value
Compression ratio	0.30 (30% expert removal)
Original experts per layer	512
Remaining experts per layer	359
Pruning method	REAP
Distance measure	Angular (cosine)
Router weight renormalization	Yes
Seed	42
Observation batch size	8
Calibration batches	128 per category

Benchmark Results

10-task lm-eval suite, 200 samples per task, tensor_parallel_size=4, vLLM eager mode:

Task	Metric	Original	REAP 0.30	Delta
ARC-Challenge	acc_norm	58.5%	61.0%	+2.5
BoolQ	acc	93.0%	90.0%	-3.0
CommonsenseQA	acc	89.0%	85.5%	-3.5
GSM8K	flexible_extract	35.0%	17.5%	-17.5
HellaSwag	acc_norm	72.0%	63.5%	-8.5
MathQA	acc_norm	60.5%	51.5%	-9.0
OpenBookQA	acc_norm	48.5%	49.5%	+1.0
PIQA	acc_norm	80.0%	79.0%	-1.0
TruthfulQA MC2	acc	60.2%	55.5%	-4.7
WinoGrande	acc	70.0%	66.0%	-4.0

Aggregate:

Overall average: 66.7% -> 61.9% (-4.8 pts)
Reasoning average: 71.4% -> 68.8% (-2.6 pts)
Math average: 47.8% -> 34.5% (-13.3 pts)

Note: GSM8K strict-match reports 0% for all variants due to an output formatting issue; flexible-extract scores are shown instead.

Architecture

Qwen3-Coder-Next uses a hybrid linear/full attention architecture with 48 layers:

Full attention every 4th layer (12 layers)
Linear attention for remaining layers (36 layers)
MoE FFN with 359 remaining experts per layer, 10 active per token
Shared expert (intermediate size 512) in every layer
Context window: 262,144 tokens
Vocab size: 151,936

Usage

Transformers

python

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "0xSero/Qwen3-Coder-57B"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    torch_dtype="auto",
    device_map="auto",
    trust_remote_code=True,
)

messages = [{"role": "user", "content": "Write a quicksort in Python."}]
text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(text, return_tensors="pt").to(model.device)
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

vLLM

bash

vllm serve 0xSero/Qwen3-Coder-57B \
    --tensor-parallel-size 4 \
    --enforce-eager \
    --gpu-memory-utilization 0.9 \
    --max-model-len 32768

Reproducing

bash

git clone https://github.com/cerebras/reap
cd reap

python -m reap.layerwise_prune \
    --model-name Qwen/Qwen3-Coder-Next \
    --dataset-name combined \
    --compression-ratio 0.30 \
    --prune-method reap \
    --seed 42 \
    --renormalize_router_weights true \
    --batch_size 8 \
    --batches_per_category 128

License & citation

License inherited from the base model.

bibtex

@misc{lasby2025reap,
  title  = {REAP the Experts: Why Pruning Prevails for One-Shot MoE Compression},
  author = {Mike Lasby and Ivan Lazarevich and Nish Sinnadurai and Sean Lie and Yani Ioannou and Vithursan Thangarasa},
  year   = {2025}, eprint = {2510.13999}, archivePrefix = {arXiv}
}

0xSero/
Qwen3-Coder-57B

Qwen3-Coder-57B

At a glance