 |
VOOZH | about |
NVIDIA
Operating mode
Use this to bias workload recommendations toward responsiveness, background autonomy, lighter serving, or multi-GPU scale-out.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
The NVIDIA H800 is a China-export-compliant variant of the H100, retaining the full Hopper compute capability — 80 GB HBM3 and Transformer Engine with FP8 — but with NVLink bandwidth cut to approximately 400 GB/s (down from H100's 900 GB/s) and FP64 performance capped at 1 TFLOPS. For single-GPU LLM inference, H800 performance is essentially identical to H100 SXM, making it highly effective for serving 70B models at FP16. The reduced NVLink bandwidth imposes a penalty for multi-GPU tensor parallelism in large training runs, which is why it was designed to be compliant. Like the A800, it was later banned under October 2023 export controls.
Beyond LLMs
What AI tasks this GPU can handle — from text generation to image and video creation.
| Capability | Status | Representative Model | Detail |
|---|---|---|---|
| LLM Chat (7B) | Runs natively | Llama 3.1 8B Q4 | — |
| LLM Coding (30B) | Runs natively | Qwen 3 30B Q4 | — |
| LLM Large (70B) |
Architecture
Hopper is NVIDIA's datacenter-focused architecture succeeding Ampere. Built on TSMC 4N, it introduces the Transformer Engine with automatic FP8/FP16 mixed-precision training, HBM3/HBM3e memory, and NVLink 4.0 for multi-GPU scaling. The H100 flagship delivers up to 3x the AI training performance of A100.
AI Relevance
The Transformer Engine automatically manages FP8 precision for optimal training speed without accuracy loss. With up to 141 GB HBM3e (H200), Hopper GPUs can hold the largest open-weight models entirely in GPU memory, making them the workhorse of AI datacenters.
Chat
SQwen 3 32B matches Chat and keeps a practical fit profile. It is a recent-generation family, which helps on current local SOTA workloads. It fits natively with comfortable headroom. Context coverage stays within the requested workload envelope. Known distribution channels: huggingface, ollama, lm-studio.
Coding
SQwen3-Coder-Next is a specialized fit for Coding. It is a recent-generation family, which helps on current local SOTA workloads. It fits natively with comfortable headroom. Context coverage stays within the requested workload envelope. Known distribution channels: huggingface, ollama, lm-studio.
Agentic Coding
SJust out of reach
High-quality models that need a bit more memory
Image & Video Generation
51 of 52 models can generate images or video on your NVIDIA H800 80GB
| Model | Max Resolution | Gen Time | Grade |
|---|---|---|---|
| SD TurboImage | 512×512 | 0ms | S |
| Stable Diffusion 1.5Image | 512×768 | 100ms | S |
| Realistic Vision v5.1Image | 512×768 | 100ms | S |
| DreamShaper 8Image | 512×768 | 100ms | S |
| LCM DreamShaper v7 |
Multi-GPU scaling
Scale out with multiple GPUs for larger models. NVLink provides 400 GB/s inter-GPU bandwidth with 12% overhead.
| Config | Effective memory | Models that fit | Est. bandwidth |
|---|---|---|---|
| 1× NVIDIA | 80 GB | 350/374 | 3,000 GB/s |
| 2× NVIDIA | 160 GB | 359/374 | 5,280 GB/s |
| 4× NVIDIA | 320 GB | 364/374 | 10,560 GB/s |
| 8× NVIDIA | 640 GB | 373/374 | 21,120 GB/s |
Model counts use default quantization at coding workload settings. Multi-GPU scaling factor: 0.88× per additional GPU.
Upgrade paths
See what you unlock with more powerful hardware
Upgrade options
Unlocks 23 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 119%.
NVLink gives this scale-out path a cleaner inter-GPU story than PCIe-only builds.
~$30,000 MSRP
Unlocks 1 additional models that do not fit on the current setup.
~$3,999 MSRP
Unlocks 1 additional models that do not fit on the current setup.
~$9,999 MSRP
Unlocks 13 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 26%.
~$20,000 MSRP
Unlocks 14 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 41%.
~$8,000 MSRP
Compare with similar
| Runs natively |
| Llama 3.1 70B Q4 |
| — |
| Image Gen (SDXL) | Runs natively | SDXL 1.0 FP16 | ~300ms per image |
| Image Gen (Flux) | Runs natively | Flux.1 Dev FP16 | ~~1.5s per image |
| Image Gen (SD 3.5) | Runs natively | SD 3.5 Large FP16 | ~~1.9s per image |
| Video Short (25f) | Runs natively | LTX Video 2B | ~300ms/frame |
| Video Long (100f) | Runs with offload | Wan Video 14B | ~900ms/frame |
Qwen3-Coder-Next is a specialized fit for Agentic Coding. It is a recent-generation family, which helps on current local SOTA workloads. It fits natively with comfortable headroom. Context coverage stays within the requested workload envelope. Known distribution channels: huggingface, ollama, lm-studio.
Reasoning
SQwen3-Coder-Next matches Reasoning and keeps a practical fit profile. It is a recent-generation family, which helps on current local SOTA workloads. It fits natively with comfortable headroom. Context coverage stays within the requested workload envelope. Known distribution channels: huggingface, ollama, lm-studio.
RAG
SQwen 3.5 27B matches RAG and keeps a practical fit profile. It is a recent-generation family, which helps on current local SOTA workloads. It fits natively with comfortable headroom. Context coverage stays within the requested workload envelope. Known distribution channels: huggingface, ollama, lm-studio.
| 512×768 |
| 0ms |
| S |
| PixArt-SigmaImage | 1024×1024 | 300ms | S |
| FramePack I2VVideo | 1280×720 | 600ms/frame | S |
| SDXL TurboImage | 512×512 | 0ms | S |
| SDXL LightningImage | 1024×1024 | 100ms | S |
| Stable Diffusion XL 1.0Image | 1024×1024 | 300ms | S |
| Playground v2.5Image | 1024×1024 | 500ms | S |
| RealVisXL v5.0Image | 1024×1024 | 400ms | S |
| DreamShaper XLImage | 1024×1024 | 400ms | S |
| Juggernaut XL v9Image | 1024×1024 | 400ms | S |
| Animagine XL 3.1Image | 1024×1024 | 400ms | S |
| Pony Diffusion V6 XLImage | 1024×1024 | 400ms | S |
| Animagine XL 4.0Image | 1024×1024 | 400ms | S |
| Illustrious XLImage | 1024×1024 | 400ms | S |
| Wan Video 2.1 1.3BVideo | 480×832 | 200ms/frame | S |
| Stable Diffusion 3.5 MediumImage | 1024×1024 | 600ms | S |
| Flux.2 Klein 4BImage | 1024×1024 | 100ms | S |
| LTX Video 2BVideo | 1280×720 | 300ms/frame | S |
| KolorsImage | 1024×1024 | 700ms | S |
| Stable CascadeImage | 1024×1024 | 900ms | S |
| AuraFlow v0.3Image | 1536×1536 | ~1.5s | S |
| Stable Diffusion 3.5 LargeImage | 1024×1024 | ~1.9s | S |
| Stable Diffusion 3.5 Large TurboImage | 1024×1024 | 300ms | S |
| CogVideoX 2BVideo | 720×480 | 300ms/frame | S |
| HunyuanVideoVideo | 720×1280 | 600ms/frame | S |
| ChromaImage | 1024×1024 | 300ms | S |
| Z-Image TurboImage | 1536×1536 | 400ms | S |
| Flux.1 DevImage | 1024×1024 | ~1.5s | S |
| Flux.1 SchnellImage | 1024×1024 | 300ms | S |
| LTX Video 13BVideo | 1280×720 | 600ms/frame | S |
| Flux.1 Kontext DevImage | 1024×1024 | ~1.7s | S |
| AnimateDiff v1.5.3Video | 512×768 | 200ms/frame | S |
| Cosmos Diffusion 7BVideo | 1024×576 | 500ms/frame | S |
| CogVideoX 5BVideo | 720×480 | 400ms/frame | S |
| Wan2.2 TI2V 5BVideo | 832×480 | 400ms/frame | S |
| Flux.2 Klein 9BImage | 1024×1024 | 200ms | S |
| Flux.1 Fill DevImage | 1024×1024 | ~1.5s | S |
| Mochi 1 PreviewVideo | 848×480 | 600ms/frame | S |
| HunyuanVideo 1.5Video | 720×1280 | 500ms/frame | S |
| Helios 14BVideo | 1280×720 | 600ms/frame | S |
| SkyReels V2 14BVideo | 1280×720 | 600ms/frame | S |
| Wan Video 2.1 14BVideo | 720×1280 | 600ms/frame | S |
| Wan Video 2.2 14BVideo | 720×1280 | 600ms/frame | S |
| Qwen ImageImage | 1024×1024 | 600ms | S |
| Qwen Image EditImage | 1024×1024 | 600ms | S |
| Flux.2 DevImage | 1024×1024 | ~16.2s | S |
| MAGI-1Video | 1280×720 | 800ms/frame | A |
| HunyuanImage 3.0Image | 256×256 | ~1s | F |
Image models estimated at 1024×1024 (28 steps, FP16). Video models estimated at 768×512 (25 frames, 30 steps, FP16). Actual performance varies with runtime and system load.
Buying advice
Excellent choice for local AI
Runs 36 of 50 top models well — a strong all-rounder for local inference.
80.0 GB
VRAM
$30,000
MSRP
$375/GB
Cost per GB VRAM
Best models for this GPU
What will limit you first
This setup is broadly balanced for this model.
No major red flags
This recommendation has enough memory headroom and acceptable estimated speed for the selected workload.
Best upgrade itinerary
Unlocks 1 additional models that do not fit on the current setup.
Want more headroom? Mac Studio M2 Ultra 128GB (128.0 GB unified memory) is the next step up.
