The Eagle3 draft model was trained using the
SpecForge
framework for the Qwen3-235B-A22B model, leveraging a combination of UltraChat and ShareGPT datasets.
Benchmark Results
gsm8k (200 questions)
Output throughput: 224.168 token/s
Accept length: 3.538
mtbench (80 questions)
Output throughput: 241.5 token/s
Accept length: 3.019
Usage
You can use this Eagle3 draft model in
SGLang
with the following command:
Qwen3-235B-A22B-EAGLE3 huggingface.co is an AI model on huggingface.co that provides Qwen3-235B-A22B-EAGLE3's model effect (), which can be used instantly with this lmsys Qwen3-235B-A22B-EAGLE3 model. huggingface.co supports a free trial of the Qwen3-235B-A22B-EAGLE3 model, and also provides paid use of the Qwen3-235B-A22B-EAGLE3. Support call Qwen3-235B-A22B-EAGLE3 model through api, including Node.js, Python, http.
Qwen3-235B-A22B-EAGLE3 huggingface.co is an online trial and call api platform, which integrates Qwen3-235B-A22B-EAGLE3's modeling effects, including api services, and provides a free online trial of Qwen3-235B-A22B-EAGLE3, you can try Qwen3-235B-A22B-EAGLE3 online for free by clicking the link below.
lmsys Qwen3-235B-A22B-EAGLE3 online free url in huggingface.co:
Qwen3-235B-A22B-EAGLE3 is an open source model from GitHub that offers a free installation service, and any user can find Qwen3-235B-A22B-EAGLE3 on GitHub to install. At the same time, huggingface.co provides the effect of Qwen3-235B-A22B-EAGLE3 install, users can directly use Qwen3-235B-A22B-EAGLE3 installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
Qwen3-235B-A22B-EAGLE3 install url in huggingface.co: