Home / Models / ZDTaichu5.0-9B

How to Run ZDTaichu5.0-9B Locally — Setup Guide (September 2026)

ZDTaichu5.0-9B pairs a Qwen3.5-9B language backbone with a C-RADIOv4-H vision encoder for general image/video understanding, spatial reasoning and agentic tool use. The language side reuses Qwen3.5-9B's hybrid attention exactly, so its GGUFs land on stock llama.cpp's existing qwen35 graph, and the whole thing — LM plus vision tower — fits under 8 GB at a useful quant.

Numbers you can trust

We don't print guessed speeds. Every measured number on this page came from TurboLLM's own auto-tuner running ZDTaichu5.0-9B on real hardware — labeled as such. When you run it yourself, TurboLLM benchmarks it on your exact GPU and shows a VRAM-fit verdict before you load, and the real measured tokens/sec once it's running.

Does it fit your GPU?

9.8B dense language backbone: 32 layers, 24 linear-attention (recurrent, no growing KV cache) and 8 full-attention (4 KV heads, 256 head dimension) — the same hybrid split as the Qwen3.5-9B base it's built on. Vision comes from a separate C-RADIOv4-H encoder shipped as its own GGUF (mmproj) file. Native context is 128K tokens.

ZDTaichu5.0-9B
TaichuAI · GGUF
Dense · hybrid gated-delta attention · vision + tools
visionthinkingtools
VRAMFit
8 GBQ4_K_M (5.63 GB) + mmproj-Q8_0 (0.88 GB) = 6.51 GB → room left for images and a full 128K context
16 GBQ8_0 (9.53 GB) + mmproj-Q8_0 (0.88 GB) = 10.41 GB → full-precision language weights with headroom to spare
24 GBBF16 (17.92 GB) + mmproj-BF16 (1.65 GB) = 19.57 GB → unquantized language and vision weights
TaichuAI/ZDTaichu5.0-9B-GGUF
View on Hugging Face →

How to run it

  1. Install TurboLLM

    npx turbollm — no install step, works on Windows, macOS, and Linux. It detects your GPU and auto-provisions a matching llama-server build (no CUDA toolkit, no Python, no compiler).

  2. Get the model

    Paste TaichuAI/ZDTaichu5.0-9B-GGUF into TurboLLM's in-app Hugging Face search — it lists every quant with a VRAM-fit verdict against your real free VRAM, and downloads are resumable and SHA-256 verified.

  3. Let auto-fit pick the setup

    Flip the auto-fit toggle and TurboLLM chooses the GPU/CPU layer split (and, for a Mixture-of-Experts model, the expert-offload split) for your card — the same manual flag-hunting every other guide walks you through by hand.

  4. Load it

    TurboLLM benchmarks the model on your exact GPU as it loads and shows the real measured tokens/sec — not a number copied from someone else's card.

  5. Use it

    Chat in the built-in UI, or point any OpenAI- or Anthropic-compatible tool (including Claude Code) at TurboLLM's local API — see the API docs.

One honest caveat

ZDTaichu5.0-9B's headline feature is entropy-gated adaptive recurrent reasoning — extra recurrent refinement steps allocated per token based on difficulty, served in the vendor's own stack through a patched vLLM build (v0.26.0-zdtaichu). Whether that per-token depth-gating survives the GGUF conversion, or the GGUF just runs a fixed-depth forward pass, isn't stated by the GGUF uploader. Vision needs the separate mmproj file from the same repo. Benchmarks on the model card are vendor-run, not independently measured, so this page ships without a numbers table.

FAQ

Will ZDTaichu5.0-9B run on my GPU?

Check the VRAM table above for the tier closest to your card. If you're not sure which tier you're in, or want picks tailored to your exact hardware, use What can I run? — enter your VRAM (or Apple unified memory) and it suggests the model, quant, and offload split that fits, the same logic TurboLLM's auto-fit runs when you load a model.

Which quant should I use for ZDTaichu5.0-9B?

Pick the highest quant that fits your VRAM with room left over for the KV cache and compute buffer — not just the model weights. As a starting point, Q4_K_M is usually the sweet spot (small quality loss for a large size drop); step up to Q5_K_M/Q6_K/Q8_0 if you have headroom, or down to Q3_K_M/an IQ quant if you don't. TurboLLM shows a VRAM-fit verdict for every quant against your card's real free VRAM before you download — see Quantization explained for what the letters and numbers actually mean, or llama.cpp flags that actually matter for the other settings worth knowing.

Why does ZDTaichu5.0-9B suddenly get slow or stutter?

The usual cause is silent VRAM spill: once a model's weights plus its KV cache stop fitting in VRAM, the driver quietly offloads the overflow to system RAM over PCIe, and that path runs 5–10× slower with no error message — just a sudden slowdown. TurboLLM's auto-fit sizes the quant and GPU/CPU split against your card's real free VRAM (including KV-cache growth at your chosen context) specifically to avoid this, and shows a VRAM-fit verdict before you load rather than after it's already crawling. Full mechanics in VRAM spill explained.

Run it now

$ npx turbollm

One command detects your GPU, provisions the right engine, and opens the UI. New here? Start with Install & first run and Quantization explained. On different hardware, see RTX 4060 or RTX 3090. Also see how to run it or how to run it.