Model Architecture: Qwen3NextForCausalLM
- Input: Text
- Output: Text
Model Optimizations:
- Weight quantization: FP8
- Activation quantization: FP8
Release Date:
Version: 1.0
Model Developers:: Red Hat

Quantized version of Qwen/Qwen3-Next-80B-A3B-Thinking.

Model Optimizations

This model was obtained by quantizing the weights and activations of Qwen/Qwen3-Next-80B-A3B-Thinking 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 of the language model are quantized.

Deployment

Use with vLLM

Initialize vLLM server:

vllm serve RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block --tensor_parallel_size 2

Send requests to the server:

python

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/Qwen3-Next-80B-A3B-Thinking-FP8-block"

messages = [
    {"role": "user", "content": "Explain quantum mechanics clearly and concisely."},
]


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

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

Creation

This model was quantized using the llm-compressor library as shown below.

Creation details

python

from datasets import load_dataset
from transformers import AutoModelForCausalLM, AutoTokenizer

from llmcompressor import oneshot
from llmcompressor.modifiers.quantization import QuantizationModifier
from llmcompressor.utils import dispatch_for_generation

# NOTE: Requires a minimum of transformers 4.57.0

MODEL_ID = "Qwen/Qwen3-Next-80B-A3B-Thinking"

# Load model.
model = AutoModelForCausalLM.from_pretrained(MODEL_ID, torch_dtype="auto")
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)


# Configure the quantization algorithm and scheme.
# In this case, we:
#   * quantize the weights to fp8 with per channel via ptq
#   * quantize the activations to fp8 with dynamic per token
recipe = QuantizationModifier(
    targets="Linear", scheme="FP8_BLOCK", ignore=[
        "lm_head",
        "re:.*mlp.gate$",
        "re:.*mlp.shared_expert_gate$",
        "re:.*linear_attn.*",
    ],
)

# Apply quantization.
oneshot(model=model, recipe=recipe)

# Confirm generations of the quantized model look sane.
print("========== SAMPLE GENERATION ==============")
dispatch_for_generation(model)
input_ids = tokenizer("Hello my name is", return_tensors="pt").input_ids.to(
    model.device
)
output = model.generate(input_ids, max_new_tokens=20)
print(tokenizer.decode(output[0]))
print("==========================================")

# Save to disk in compressed-tensors format.
SAVE_DIR = MODEL_ID.rstrip("/").split("/")[-1] + "-FP8-block"
model.save_pretrained(SAVE_DIR)
tokenizer.save_pretrained(SAVE_DIR)

Evaluation

The model was evaluated on the OpenLLM leaderboard task, using lm-evaluation-harness. vLLM was used for all evaluations.

Evaluation details

Openllm V1

lm_eval \
  --model vllm \
  --model_args pretrained="RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block",dtype=auto,add_bos_token=True,max_model_len=16384,tensor_parallel_size=2,gpu_memory_utilization=0.9,enable_chunked_prefill=True,trust_remote_code=True \
  --tasks openllm \
  --write_out \
  --batch_size auto \
  --show_config

Openllm V2

lm_eval \
  --model vllm \
  --model_args pretrained="RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block",dtype=auto,add_bos_token=False,max_model_len=16384,tensor_parallel_size=2,gpu_memory_utilization=0.7,disable_log_stats=True,enable_chunked_prefill=True,trust_remote_code=True \
  --tasks leaderboard \
  --apply_chat_template \
  --fewshot_as_multiturn \
  --write_out \
  --batch_size auto \
  --show_config

Coding Benchmarks

evalplus.evaluate --model "RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block" \
                  --dataset "humaneval" \
                  --backend vllm \
                  --tp 2 \
                  --greedy

evalplus.evaluate --model "RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block" \
                --dataset "mbpp" \
                --backend vllm \
                --tp 2 \
                --greedy

Accuracy

Benchmark	Qwen/Qwen3-Next-80B-A3B-Thinking	Qwen/Qwen3-Next-80B-A3B-Thinking-FP8	Recovery (%)	RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-block	Recovery (%)	RedHatAI/Qwen3-Next-80B-A3B-Thinking-FP8-dynamic	Recovery (%)	RedHatAI/Qwen3-Next-80B-A3B-Thinking-quantized.w4a16	Recovery (%)
AIME 2025	60.00	50.00	83.33	50.00	83.33	50.00	83.33	50.00	83.33
MATH-500	94.00	87.60	93.19	94.40	100.43	86.80	92.34	87.20	92.77
GPQA Diamond	73.74	76.77	104.11	76.77	104.11	74.75	101.37	73.74	100.00