Published: September 10, 2026 | Reading time: 10 minutes
On September 3, 2026, researchers from Duke University and Tsinghua University published a groundbreaking paper that answers a fundamental question in AI code generation:
Can language models write code using diffusion models — and do it in just one step?
The answer is yes.
PlaidQ is a continuous (Gaussian) latent-diffusion language model that works differently from traditional autoregressive models:
The key innovation is distillation — reducing the number of denoising steps:
| Steps | Description | Performance |
|---|---|---|
| 512 | Original diffusion process | Baseline |
| 16 | Distilled to 16 steps | Student outperforms teacher on HumanEval pass@10 |
| 1 | Distilled to single step | Can generate executable code, but HumanEval pass@1 is only 7.07 |
The 16-step model demonstrates that students can surpass teachers on certain benchmarks. The 1-step model shows the feasibility of parallel generation, though it's not yet reliable for high-quality coding.
import torch
from plaidq import PlaidQModel
model = PlaidQModel.from_pretrained("plaidq-0.7b")
prompt = """
def fibonacci(n):
"""Generate Fibonacci sequence."""
if n <= 1:
return n
return fibonacci(n-1) + fibonacci(n-2)
"""
result = model.generate(prompt, num_steps=16)
print(result)
result_one_step = model.generate(prompt, num_steps=1)
print(result_one_step)
| Model | HumanEval pass@1 | HumanEval pass@10 | MBPP pass@1 |
|---|---|---|---|
| Teacher (512 steps) | 65.85 | 78.05 | 72.30 |
| Student (16 steps) | 63.41 | 80.73 | 70.15 |
| Student (1 step) | 7.07 | 15.30 | 8.20 |
Key Insight: The 16-step student outperforms the teacher on HumanEval pass@10, demonstrating effective distillation. The 1-step model shows feasibility but needs improvement.
from plaidq.distill import distill_model
teacher = PlaidQModel.from_pretrained("plaidq-teacher")
student = distill_model(
teacher=teacher,
num_steps=16,
dataset="humaneval",
epochs=10
)
score = student.evaluate("humaneval", metric="pass@10")
print(f"Student pass@10: {score}")
student_1step = distill_model(
teacher=student,
num_steps=1,
dataset="humaneval",
epochs=5
)
PlaidQ represents a significant step toward parallel code generation using diffusion models. The ability to generate code in just 16 steps — or even 1 step — challenges the traditional autoregressive paradigm and opens new possibilities for AI code generation.
While the 1-step model is not yet reliable for production use, the 16-step model demonstrates that students can surpass teachers through effective distillation.
This research highlights the potential of diffusion-based language models and the importance of distillation in achieving high-quality, efficient code generation.
This article is based on research published by Duke University and Tsinghua University on September 3, 2026. Paper: arXiv:2609.04531 | Code: github.com/pengzhangzhi/plaidq