Qwen3.8-Flash-Next-GSQ-RCO-GGUF not loading.

[DarkReaperBoy]

Before you start

• I have read the FAQ and my problem is not answered there.
• I have read the Roadmap and this is not already planned there.
• I have searched existing issues and found no duplicate.
• I am on the latest release, or on a freshly rebuilt main when building from source.

What happened

Qwen3.8-Flash-Next-GSQ-RCO-GGUF 2 bit and 3 bit didn't load.

How did you install FreeToken

pip / uv wheel

FreeToken version

freetoken version 0.1.3

OS

Other Linux

OS details

NixOS 26.11 (Zokor) x86_64, Linux 7.2.6, Python 3.13.14

GPU and driver

NVIDIA GeForce RTX 2060 Mobile 6GB

CPU and system RAM

Intel(R) Core(TM) i7-10750H (12) @ 5.00 GHz

Checkpoint

ISTA-DASLab/Qwen3.8-Flash-Next-GSQ-RCO-GGUF

Command

ft serve --model /run/media/nako/External/ai/gguf/Qwen3.8/Qwen3.8-Flash-Next/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00001-of-00002.gguf --text-model-only

Full log

(freetoken) ╰─❮ ft serve --model /run/media/nako/External/ai/gguf/Qwen3.8/Qwen3.8-Flash-Next/Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00001-of-
00002.gguf --text-model-only
 
Traceback (most recent call last):
  File "/run/media/nako/External/ai/freetoken/.venv/bin/ft", line 10, in <module>
    sys.exit(main())
             ~~~~^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/cli.py", line 126, in main
    return runner(args[1:])
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/cli.py", line 30, in _run_serve
    launch_server(argv=argv, prog="ft serve")
    ~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/server/launch.py", line 133, in launch_server
    server_args, run_shell = parse_args(
                             ~~~~~~~~~~^
        sys.argv[1:] if argv is None else argv,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
        run_shell,
        ^^^^^^^^^^
        prog=prog,
        ^^^^^^^^^^
    )
    ^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/server/args.py", line 866, in parse_args
    cfg = cached_load_hf_config(kwargs["model_path"]).to_dict()
          ~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/utils/hf.py", line 247, in cached_load_hf_config
    return build_gguf_shim(gguf_src)
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/models/gguf/config.py", line 62, in build_gguf_shim
    arch = gguf_architecture(model_path)
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/models/gguf/reader.py", line 137, in gguf_architecture
    arch = _field_value(_reader(model_path), "general.architecture")
                        ~~~~~~~^^^^^^^^^^^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/freetoken/models/gguf/reader.py", line 126, in _reader
    return gguf.GGUFReader(model_path)
           ~~~~~~~~~~~~~~~^^^^^^^^^^^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/gguf/gguf_reader.py", line 185, in __init__
    self._build_tensors(offs, tensors_fields)
    ~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^
  File "/run/media/nako/External/ai/freetoken/.venv/lib/python3.13/site-packages/gguf/gguf_reader.py", line 328, in _build_tensors
    ggml_type = GGMLQuantizationType(raw_dtype[0])
  File "/home/nako/.local/share/uv/python/cpython-3.13.14-linux-x86_64-gnu/lib/python3.13/enum.py", line 728, in __call__
    return cls.__new__(cls, value)
           ~~~~~~~~~~~^^^^^^^^^^^^
  File "/home/nako/.local/share/uv/python/cpython-3.13.14-linux-x86_64-gnu/lib/python3.13/enum.py", line 1205, in __new__
    raise ve_exc
ValueError: np.uint32(42) is not a valid GGMLQuantizationType

Anything else

No response

[DarkReaperBoy]

wait wait, just saw the roadmap gguf is not supported, then wtf is this:
https://www.reddit.com/r/LocalLLaMA/comments/1wi46on/qwen38_flash_on_12gb_vram_15_tokenss/