ZDTaichu5.0 9B — Hardware Requirements & GPU Compatibility
VisionFunctions
ZDTaichu5.0-9B is TaichuAI's 9.8-billion-parameter multimodal foundation model, combining a Qwen3.5-9B language backbone with an NVIDIA C-RADIOv4-H vision encoder to accept text, images, and video at any resolution. Beyond general image, document, and OCR understanding, it is built for spatial reasoning (2D/3D relations, viewpoint and depth, affordances), embodied-AI planning, and multi-step agentic tool use, using an "Entropy-Gated Adaptive Recurrent Reasoning" mechanism that allocates extra recurrent computation to harder tokens. The card reports it leading spatial and agentic benchmarks among comparable 10B-scale open vision-language models. At under 10 billion parameters, it fits on a single consumer GPU once quantized. Context length is 131,072 tokens. It is released under the NVIDIA Open Model License Agreement, with the underlying Qwen3.5 component retaining its Apache 2.0 license, and was published in September 2026.
6.9K downloads
761 likesSpecifications
- Publisher
- TaichuAI
- Parameters
- 9.8B
- Architecture
- ZDTaichu5_0_ForConditionalGeneration
- Release Date
- 2026-09-04
Get Started
HuggingFace
How Much VRAM Does ZDTaichu5.0 9B Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 4.6 GB | — | 4.16 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 4.7 GB | — | 4.28 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 5.3 GB | — | 4.77 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 5.4 GB | — | 4.90 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 6.5 GB | — | 5.88 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 7.7 GB | — | 6.98 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 8.9 GB | — | 8.08 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 10.8 GB | — | 9.79 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run ZDTaichu5.0 9B?
Q4_K_M · 6.5 GBZDTaichu5.0 9B (Q4_K_M) requires 6.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 9+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run ZDTaichu5.0 9B?
Q4_K_M · 6.5 GB58 devices with unified memory can run ZDTaichu5.0 9B, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, MacBook Air 13" M3 (8 GB).
Runs great
— Plenty of headroomWhere to Download ZDTaichu5.0 9B
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Frequently Asked Questions
- How much VRAM does ZDTaichu5.0 9B need?
ZDTaichu5.0 9B requires 6.5 GB of VRAM at Q4_K_M, or 21.6 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 9.8B × 4.8 bits ÷ 8 = 5.9 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
Q4_K_M6.5 GB- What's the best quantization for ZDTaichu5.0 9B?
For ZDTaichu5.0 9B, Q4_K_M (6.5 GB) offers the best balance of quality and VRAM usage. Q5_K_S (7.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XS at 3.2 GB.
VRAM requirement by quantization
IQ2_XS3.2 GBIQ3_S4.6 GBQ3_K_L5.5 GBQ4_K_M ★6.5 GBQ5_K_S7.4 GBBF1621.6 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run ZDTaichu5.0 9B on a Mac?
ZDTaichu5.0 9B requires at least 3.2 GB at IQ2_XS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run ZDTaichu5.0 9B locally?
Yes — ZDTaichu5.0 9B can run locally on consumer hardware. At Q4_K_M quantization it needs 6.5 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is ZDTaichu5.0 9B?
At Q4_K_M, ZDTaichu5.0 9B can reach ~743 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~101 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 6.5 × 0.65 = ~805 tok/s
Estimated speed at Q4_K_M (6.5 GB)
~805 tok/s~101 tok/s~805 tok/s~743 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of ZDTaichu5.0 9B?
At Q4_K_M, the download is about 5.88 GB. The full-precision BF16 version is 19.59 GB. The smallest option (IQ2_XS) is 2.94 GB.
- Which GPUs can run ZDTaichu5.0 9B?
52 consumer GPUs can run ZDTaichu5.0 9B at Q4_K_M (6.5 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.
- Which devices can run ZDTaichu5.0 9B?
59 devices with unified memory can run ZDTaichu5.0 9B at Q4_K_M (6.5 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.