Qwen3-30B-A3B-Instruct-2507-speculator.eagle3

Qwen3-8B-speculator.eagle3

Model Overview

• Verifier: Qwen3-30B-A3B-Instruct-2507
• Speculative Decoding Algorithm: EAGLE-3
• Model Architecture: Eagle3Speculator
• Release Date: 12/12/2025
• Version: 1.0
• Model Developers: RedHat

This is a speculator model designed for use with Qwen3-30B-A3B-Instruct-2507 , based on the EAGLE-3 speculative decoding algorithm. It was trained using the speculators library on a combination of the Magpie-Align/Magpie-Pro-300K-Filtered and the HuggingFaceH4/ultrachat_200k datasets. The model was trained with thinking turned disabled. This model should be used with the Qwen3-30B-A3B-Instruct-2507 chat template, specifically through the /chat/completions endpoint.

Use with vLLM

vllm serve Qwen3-30B-A3B-Instruct-2507 \
  -tp 1 \
  --speculative-config '{
    "model": "RedHatAI/Qwen3-30B-A3B-Instruct-2507-speculator.eagle3",
    "num_speculative_tokens": 3,
    "method": "eagle3"
  }'

Evaluations

<h3>Use cases</h3> <table> <thead> <tr> <th>Use Case</th> <th>Dataset</th> <th>Number of Samples</th> </tr> </thead> <tbody> <tr> <td>Coding</td> <td>HumanEval</td> <td>168</td> </tr> <tr> <td>Math Reasoning</td> <td>gsm8k</td> <td>80</td> </tr> <tr> <td>Text Summarization</td> <td>CNN/Daily Mail</td> <td>80</td> </tr> </tbody> </table> <h3>Acceptance lengths</h3> <table> <thead> <tr> <th>Use Case</th> <th>k=1</th> <th>k=2</th> <th>k=3</th> <th>k=4</th> <th>k=5</th> </tr> </thead> <tbody> <tr> <td>Coding</td> <td>1.81</td> <td>2.47</td> <td>2.85</td> <td>3.13</td> <td>3.50</td> </tr> <tr> <td>Math Reasoning</td> <td>1.87</td> <td>2.53</td> <td>3.09</td> <td>3.57</td> <td>3.79</td> </tr> <tr> <td>Text Summarization</td> <td>1.58</td> <td>1.90</td> <td>2.08</td> <td>2.16</td> <td>2.24</td> </tr> </tbody> </table> <details> <summary>Details</summary> <strong>Configuration</strong>