Skip to main content

supported_models

Released Models

We list the models released by the SpecForge and several industrial partners below. These models are released as part of the SpecBundle models, which are trained on large-scale multi-domain datasets and deliver exceptional performance on various benchmarks.

We also include some of the models previously trained by the SpecForge team but not technically part of the SpecBundle release. We mark models trained on ShareGPT+Ultrachat datasets with a * mark and models trained on Perfect-Blend datasets but released before SpecBundle with + mark.

Llama Series

Target ModelEAGLE3 Draft Model
meta-llama/Llama-3.1-8B-Instruct🤗 Hugging Face
meta-llama/Llama-3.3-70B-Instruct🤗 Hugging Face
meta-llama/Llama-4-Scout-17B-16E-Instruct🤗 Hugging Face
meta-llama/Llama-4-Maverick-17B-128E-Instruct🤗 Hugging Face *

Qwen Series

Target ModelEAGLE3 Draft Model
Qwen/Qwen3-30B-A3B-Instruct-2507🤗 Hugging Face
Qwen/Qwen3-235B-A22B-Instruct-2507🤗 Hugging Face
Qwen/Qwen3-Next-80B-A3B-Instruct-FP8🤗 Hugging Face

Qwen Coder Series

Target ModelEAGLE3 Draft Model
Qwen/Qwen3-Coder-30B-A3B-Instruct🤗 Hugging Face
Qwen/Qwen3-Coder-480B-A35B-Instruct🤗 Hugging Face

Ling Series

Target ModelEAGLE3 Draft Model
inclusionAI/Ling-flash-2.0🤗 Hugging Face

Kimi Series

Target ModelEAGLE3 Draft Model
moonshotai/Kimi-K2-Instruct🤗 Hugging Face

GPT-OSS Series

Target ModelEAGLE3 Draft Model
openai/gpt-oss-20b🤗 Hugging Face +
openai/gpt-oss-120b🤗 Hugging Face +

Nex Series

Target ModelEAGLE3 Draft Model
nex-agi/Qwen3-30B-A3B-Nex-N1🤗 Hugging Face
nex-agi/Qwen3-32B-Nex-N1🤗 Hugging Face