What it does
Core capabilities at a glance
- Eagle
- GPU
- LLM Inference
- Sglang
- Speculative Decoding
- Tokenspeed
- Vllm
Deep dive
The full breakdown - performance, comparisons, and setup
EAGLE-4
EAGLE-4 is a local inference server - Official Inference Code of EAGLE-4.
Overview
Paper (EAGLE-4) · EAGLE-4 draft for Qwen3.8-27B · Documentation
Figure 1(a) of the EAGLE-4 paper: Qwen3.8-27B on the SPEED-Bench qualitative split, 4× NVIDIA B300, one request, temperature 0. The numbers this release measured with EAGLE-4 are in the support matrix .
Leave speculation on. EAGLE-4 adapts to your hardware, batch and workload.
For high load, serve SGLang on four B300 GPUs with the high-load recipe, which keeps pace with plain decoding up to 512 concurrent requests.
EAGLE-4 (Extrapolation Algorithm for Greater Language-model Efficiency) speeds up LLM inference with speculative decoding. EAGLE-4 draft models are built to run with 'eagle-harness', which decides how far to speculate from measured costs on your setup, so every draft tree fits the batch size, the hardware and the workload. The output stays exactly what the target model would generate on its own: up to 3.4× faster generation on Qwen3.8-27B, with the same answers. Measured speedups
This repository holds 'eagle-harness', the package that brings EAGLE-4 to SGLang, vLLM and TokenSpeed. Install it, point your engine at an EAGLE-4 draft model, and it serves with EAGLE-4.
EAGLE-4 is open-source, written primarily in Python, with 1,041 GitHub stars under the Other license. The latest release is v0.1.4 (2026-10-28).
Key capabilities
From the project's documentation:
- Install: pins, wheel indexes, Docker images and the ROCm closure.
- Engine guides with the measured recipes: SGLang, vLLM,
- Support matrix and known limits.
- How it works, kernels, compatibility,
- SPEED-Bench is by NVIDIA. Model weights are licensed by their owners.
Install
A quick way to get started (always check the official docs for the latest):
pip install "eagle-harness[sglang]" "sglang==0.5.23.dev252+eagle.9" \How it fits a local-AI stack
EAGLE-4 runs on your own hardware, so pair it with a model and a GPU sized to your needs. Use the VRAM calculator to pick a model that fits your card, and see what you can run for hardware guidance. Related local inference servers in the directory:
Sources
- Source code & docs: eagle-hq/EAGLE-4
- Official website: https://eagle.ai
Stats from GitHub, 2026-10-30.
Frequently asked
Quick answers to common questions
What is EAGLE-4?
EAGLE-4 is a inference-server tool for local AI workloads. Official Inference Code of EAGLE-4
Is EAGLE-4 free and open source?
Yes, EAGLE-4 has 1,041 GitHub stars and is licensed under Other. You can self-host it for free on linux, docker.
What platforms does EAGLE-4 support?
EAGLE-4 runs on linux, docker.
What hardware do I need for EAGLE-4?
The hardware requirements depend on which models you run. Check our hardware directory for compatible GPUs and systems. EAGLE-4 has 1,041 GitHub stars and an active community.
Does EAGLE-4 support GPU acceleration?
EAGLE-4 supports GPU acceleration via CUDA, Metal, or Vulkan depending on your platform. For the best performance, pair it with an NVIDIA RTX 4090 or 5090.
What are the best alternatives to EAGLE-4?
Popular alternatives include other inference-server tools in our directory. Browse our full collection at /tool for comparisons, community reviews, and benchmark data to find the right fit for your workflow.
How much does EAGLE-4 cost?
EAGLE-4 is free-open-source. It is completely free and open source to self-host.
Pairs well with
Complementary tools, models, and hardware