Apertus-8B-Instruct-2509-FP8-dynamic 👁 Model Icon

👁 Validated Badge

Model Overview

• Model Architecture: ApertusForCausalLM
  • Input: Text
  • Output: Text
• Model Optimizations:
  • Weight quantization: FP8
  • Activation quantization: FP8
• Release Date: 9/18/2025
• Version: 1.0
• Model Developers: Red Hat
• ModelCar Storage URI: oci://registry.redhat.io/rhai/modelcar-apertus-8b-instruct-2509-fp8-dynamic:3.0
• Validated on RHOAI 3.2: quay.io/modh/vllm:rhoai-3.2-cuda
• Validated on RHAIIS 3.2.5: http://registry.redhat.io/rhaiis/vllm-cuda-rhel9:3.2.5-1766067105
• Validated on vLLM: 0.11.2

Quantized version of swiss-ai/Apertus-8B-2509.

Model Optimizations

This model was obtained by quantizing the weights and activations of swiss-ai/Apertus-8B-2509 to FP8 data type. This optimization reduces the number of bits per parameter from 16 to 8, reducing the disk size and GPU memory requirements by approximately 50%. Only the weights and activations of the linear operators within transformers blocks are quantized.

Deployment

Use with vLLM

1. Initialize vLLM server:

vllm serve RedHatAI/Apertus-8B-Instruct-2509-FP8-dynamic

2. Send requests to the server:

from openai import OpenAI

# Modify OpenAI's API key and API base to use vLLM's API server.
openai_api_key = "EMPTY"
openai_api_base = "http://<your-server-host>:8000/v1"

client = OpenAI(
 api_key=openai_api_key,
 base_url=openai_api_base,
)

model = "RedHatAI/Apertus-8B-Instruct-2509-FP8-dynamic"

messages = [
 {"role": "user", "content": "Give me a short introduction to large language model."},
]

outputs = client.chat.completions.create(
 model=model,
 messages=messages,
)

generated_text = outputs.choices[0].message.content
print(generated_text)

Creation

This model was created with llm-compressor by running the code snippet below.

Evaluation

The model was evaluated on OpenLLM Leaderboard V1, using the following command:

Accuracy

Category	Metric	swiss-ai/Apertus-8B-Instruct-2509	RedHatAI/Apertus-8B-Instruct-2509-FP8-dynamic	Recovery (%)
OpenLLM V1	ARC-Challenge (Acc-Norm, 25-shot)	65.02	65.59	101.4
	GSM8K (Strict-Match, 5-shot)	58.07	55.50	95.6
	HellaSwag (Acc-Norm, 10-shot)	80.87	81.06	100.2
	MMLU (Acc, 5-shot)	61.97	61.86	99.8
	TruthfulQA (MC2, 0-shot)	58.14	58.18	100.1
	Winogrande (Acc, 5-shot)	75.14	75.45	100.4
	Average Score	66.54	66.33	99.7

Base model

swiss-ai/Apertus-8B-2509

swiss-ai/Apertus-8B-Instruct-2509

(34)

January 2026 Collection of third-party generative AI models validated by Red Hat AI for use across the Red Hat AI Product Portfolio. • 4 items • Updated Apr 30 • 2