[Bug] Qwen3-Next run failed.

[jiapingW]

Checklist

• I searched related issues but found no solution.
• The bug persists in the latest version.
• Issues without environment info and a minimal reproducible demo are hard to resolve and may receive no feedback.
• If this is not a bug report but a general question, please start a discussion at https://github.com/sgl-project/sglang/discussions. Otherwise, it will be closed.
• Please use English. Otherwise, it will be closed.

Describe the bug

Qwen3-Next run failed.

Reproduction

I run the latest code use the following command:

python -m sglang.launch_server \
  --model Qwen/Qwen3-Next-80B-A3B-Instruct \
  --tp 8 \
  --host 0.0.0.0 \
  --port 8000

The error message is

[2025-12-06 02:08:52 TP7] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP0] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP2] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP1] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP5] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP3] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP4] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

[2025-12-06 02:08:52 TP6] Scheduler hit an exception: Traceback (most recent call last):
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 2656, in run_scheduler_process
    scheduler = Scheduler(
                ^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 430, in __init__
    self.init_cache_with_memory_pool()
  File "/home/jiapingW/sglang/python/sglang/srt/managers/scheduler.py", line 771, in init_cache_with_memory_pool
    self.tree_cache = MambaRadixCache(params)
                      ^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jiapingW/sglang/python/sglang/srt/mem_cache/mamba_radix_cache.py", line 324, in __init__
    assert isinstance(params.token_to_kv_pool_allocator, TokenToKVPoolAllocator)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AssertionError

Environment

Python: 3.12.3 (main, Nov  6 2025, 13:44:16) [GCC 13.3.0]
CUDA available: True
GPU 0,1,2,3,4,5,6,7: NVIDIA B200
GPU 0,1,2,3,4,5,6,7 Compute Capability: 10.0
CUDA_HOME: /usr/local/cuda
NVCC: Cuda compilation tools, release 12.8, V12.8.93
CUDA Driver Version: 580.95.05
PyTorch: 2.9.1+cu128
sglang: 0.5.6
sgl_kernel: 0.3.18.post2
flashinfer_python: 0.5.3
flashinfer_cubin: 0.5.3
flashinfer_jit_cache: Module Not Found
triton: 3.5.1
transformers: 4.57.1
torchao: 0.9.0
numpy: 2.3.5
aiohttp: 3.13.2
fastapi: 0.123.10
hf_transfer: 0.1.9
huggingface_hub: 0.36.0
interegular: 0.3.3
modelscope: 1.32.0
orjson: 3.11.4
outlines: 0.1.11
packaging: 25.0
psutil: 7.1.3
pydantic: 2.12.5
python-multipart: 0.0.20
pyzmq: 27.1.0
uvicorn: 0.38.0
uvloop: 0.22.1
vllm: Module Not Found
xgrammar: 0.1.27
openai: 2.6.1
tiktoken: 0.12.0
anthropic: 0.75.0
litellm: Module Not Found
decord2: 2.0.0
NVIDIA Topology: 
        GPU0    GPU1    GPU2    GPU3    GPU4    GPU5    GPU6    GPU7    CPU Affinity    NUMA Affinity   GPU NUMA ID
GPU0     X      NV18    NV18    NV18    NV18    NV18    NV18    NV18    0-239   0               N/A
GPU1    NV18     X      NV18    NV18    NV18    NV18    NV18    NV18    0-239   0               N/A
GPU2    NV18    NV18     X      NV18    NV18    NV18    NV18    NV18    0-239   0               N/A
GPU3    NV18    NV18    NV18     X      NV18    NV18    NV18    NV18    0-239   0               N/A
GPU4    NV18    NV18    NV18    NV18     X      NV18    NV18    NV18    0-239   0               N/A
GPU5    NV18    NV18    NV18    NV18    NV18     X      NV18    NV18    0-239   0               N/A
GPU6    NV18    NV18    NV18    NV18    NV18    NV18     X      NV18    0-239   0               N/A
GPU7    NV18    NV18    NV18    NV18    NV18    NV18    NV18     X      0-239   0               N/A

Legend:

  X    = Self
  SYS  = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
  NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
  PHB  = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
  PXB  = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
  PIX  = Connection traversing at most a single PCIe bridge
  NV#  = Connection traversing a bonded set of # NVLinks

Hypervisor vendor:: KVM
ulimit soft: 1048576

[Chen-0210]

Could you share how you start the server?

[jiapingW]

I have updated the command.

Could you share how you start the server?

[Chen-0210]

I couldn’t reproduce it on tp4. Since I don’t have an 8 GPU machine, could you check whether the issue still exists on tp4? @jiapingW

[jiapingW]

I couldn’t reproduce it on tp4. Since I don’t have an 8 GPU machine, could you check whether the issue still exists on tp4? @jiapingW

Yes,it's also exist on tp4. You need to use the lasted code to test rather than sglang0.5.6.

[Chen-0210]

It only happens on Blackwell GPUs. You can temporarily try --attention-backend triton or --disable-radix-cache to avoid it.

[jiapingW]

It only happens on Blackwell GPUs. You can temporarily try --attention-backend triton or --disable-radix-cache to avoid it.
It works well, thanks.

[Chen-0210]

Could you please reopen this issue and keep it open until #14560 is merged?