Intel
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.
Intel Gaudi 3 is a dedicated AI accelerator — not a traditional GPU — designed from the ground up for large-scale LLM training and inference. It delivers 1.8 PFlops of BF16/FP8 compute and 3.7 TB/s of HBM2e bandwidth across 128 GB of on-package memory. Intel claims Gaudi 3 outperforms the NVIDIA H100 by 50% on average inference throughput for models like Llama 7B, 70B, and Falcon 180B, while delivering 40% better inference power efficiency. It integrates natively with PyTorch, Hugging Face, DeepSpeed, and vLLM via Intel's Synapse AI software stack.
Beyond LLMs
What AI tasks this GPU can handle — from text generation to image and video creation.
| Capability | Status | Representative Model |
|---|---|---|
| LLM Chat (7B) | Runs natively | Llama 3.1 8B Q4 |
| LLM Coding (30B) | Runs natively | Qwen 3 30B Q4 |
| LLM Large (70B) | Runs natively | Llama 3.1 70B Q4 |
| Image Gen (SDXL) | Runs natively | SDXL 1.0 FP16 |
| Image Gen (Flux) | Runs natively | Flux.1 Dev FP16 |
| Image Gen (SD 3.5) | Runs natively | SD 3.5 Large FP16 |
| Video Short (25f) | Runs natively | LTX Video 2B |
| Video Long (100f) | Runs natively | Wan Video 14B |
Architecture
Gaudi is Intel's purpose-built AI training and inference accelerator (acquired from Habana Labs). Gaudi 3 features 128 GB HBM2e and a dedicated Matrix Math Engine designed specifically for transformer workloads.
AI Relevance
Purpose-built for AI with integrated networking (24x 200GbE) for multi-node scaling. Gaudi 3 targets direct competition with NVIDIA H100 for transformer training and inference, with competitive TCO claims.
Buying advice
Excellent choice for local AI
Runs 36 of 50 top models well — a strong all-rounder for local inference.
128.0 GB
VRAM
$15,000
MSRP
$117/GB
Cost per GB VRAM
Best models for this GPU
What will limit you first
The raw memory story may look fine, but the software ecosystem is still a constraint here.
Runtime ecosystem is narrower than CUDA
Intel GPUs can look attractive on memory per dollar, but local AI tooling, kernels, and model coverage are still broader and easier on CUDA today.
Best upgrade itinerary
Prefer CUDA if you want the path of least resistance
If your goal is maximum runtime coverage, easier troubleshooting, and better support for new local AI releases, CUDA is usually still the safer upgrade path.
Unlocks 2 additional models that do not fit on the current setup.
Want more headroom? NVIDIA H200 141GB (141.0 GB VRAM) is the next step up.
Chat
SThis model is a direct match for chat. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, lm-studio.
Coding
SThis model is a direct match for coding. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.
Agentic Coding
SThis model is still usable for agentic-coding, but it is not the most specialized pick. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, lm-studio.
Reasoning
SThis model is a direct match for reasoning. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, lm-studio.
RAG
SThis model is a direct match for rag. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, lm-studio.
Qwen 3.5 122B A10B
Mistral Small 4 119BDevstral 2 123B Instruct
GPT-OSS 120B
Command A 111BPixtral Large 124BLeanstral 119B A6BQwen3-Coder-NextQwen 2.5 VL 72BQwen 3.6 35B A3BQwen3-Coder 30B A3B InstructQwen 3.5 27BQwen3-VL 30B A3B InstructQwen 3.5 35B A3BQwen 3.6 27BQwen 3 32BMagistral Small 2507Devstral Small 2 24B InstructQwen 3 30B A3B
Nemotron 3 Nano 30BQwen 3.5 9BQwen 3 14B
Gemma 4 31BDevstral Small 1.1
Phi-4-reasoning-plus 14BNemotron Cascade 2 30B A3BQwen 3 8BGPT-OSS 20BQwen 3.5 4B
EXAONE 4.0 32BGemma 4 26B A4BMinistral 3 14BNemotron Nano 8BPhi-4 Mini Reasoning 4B
Jina Embeddings v3
BGE M3Qwen 3.5 397B A17B
Kimi K2.5Kimi K2.6
DeepSeek V4 ProDeepSeek V4 Flash
GLM-5.1GLM-5DeepSeek V3.2Qwen 3 235B A22BQwen3-Coder 480B A35B InstructDeepSeek Coder V2 236BDeepSeek R1 671BDeepSeek V3.1 671BJust out of reach
High-quality models that need a bit more memory
Image & Video Generation
52 of 52 models can generate images or video on your Gaudi 3 128GB
| 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 v7Image | 512×768 | 0ms | S |
| PixArt-SigmaImage | 1024×1024 | 400ms | S |
| FramePack I2VVideo | 1280×720 | 700ms/frame | S |
| SDXL TurboImage | 512×512 | 0ms | S |
| SDXL LightningImage | 1024×1024 | 100ms | S |
| Stable Diffusion XL 1.0Image | 1024×1024 | 400ms | S |
| Playground v2.5Image | 1024×1024 | 600ms | 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 | 300ms/frame | S |
| Stable Diffusion 3.5 MediumImage | 1024×1024 | 700ms | S |
| Flux.2 Klein 4BImage | 1024×1024 | 100ms | S |
| LTX Video 2BVideo | 1280×720 | 300ms/frame | S |
| KolorsImage | 1024×1024 | 800ms | S |
| Stable CascadeImage | 1024×1024 | ~1s | S |
| AuraFlow v0.3Image | 1536×1536 | ~1.8s | S |
| Stable Diffusion 3.5 LargeImage | 1024×1024 | ~2.2s | S |
| Stable Diffusion 3.5 Large TurboImage | 1024×1024 | 400ms | S |
| CogVideoX 2BVideo | 720×480 | 300ms/frame | S |
| HunyuanVideoVideo | 720×1280 | 700ms/frame | S |
| ChromaImage | 1024×1024 | 400ms | S |
| Z-Image TurboImage | 1536×1536 | 400ms | S |
| Flux.1 DevImage | 1024×1024 | ~1.8s | S |
| Flux.1 SchnellImage | 1024×1024 | 300ms | S |
| LTX Video 13BVideo | 1280×720 | 700ms/frame | S |
| Flux.1 Kontext DevImage | 1024×1024 | ~2s | S |
| AnimateDiff v1.5.3Video | 512×768 | 200ms/frame | S |
| Cosmos Diffusion 7BVideo | 1024×576 | 600ms/frame | S |
| CogVideoX 5BVideo | 720×480 | 500ms/frame | S |
| Wan2.2 TI2V 5BVideo | 832×480 | 500ms/frame | S |
| Flux.2 Klein 9BImage | 1024×1024 | 200ms | S |
| Flux.1 Fill DevImage | 1024×1024 | ~1.7s | S |
| Mochi 1 PreviewVideo | 848×480 | 700ms/frame | S |
| HunyuanVideo 1.5Video | 720×1280 | 600ms/frame | S |
| Helios 14BVideo | 1280×720 | 700ms/frame | S |
| SkyReels V2 14BVideo | 1280×720 | 700ms/frame | S |
| Wan Video 2.1 14BVideo | 720×1280 | 700ms/frame | S |
| Wan Video 2.2 14BVideo | 720×1280 | 700ms/frame | S |
| Qwen ImageImage | 1024×1024 | 700ms | S |
| Qwen Image EditImage | 1024×1024 | 700ms | S |
| Flux.2 DevImage | 1024×1024 | ~18.7s | S |
| MAGI-1Video | 1280×720 | 900ms/frame | S |
| HunyuanImage 3.0Image | 256×256 | ~1.2s | D |
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.
Multi-GPU scaling
Scale out with multiple GPUs for larger models. PCIe interconnect with 15% scaling overhead.
| Config | Effective memory | Models that fit | Est. bandwidth |
|---|---|---|---|
| 1× Gaudi | 128 GB | 351/374 | 3,700 GB/s |
| 2× Gaudi | 256 GB | 363/374 | 6,290 GB/s |
| 4× Gaudi | 512 GB | 371/374 | 12,580 GB/s |
| 8× Gaudi | 1024 GB | 374/374 | 25,160 GB/s |
Model counts use default quantization at coding workload settings. Multi-GPU scaling factor: 0.85× per additional GPU.
Upgrade paths
See what you unlock with more powerful hardware
Upgrade options
8× Gaudi 3 128GBMulti-GPUUnlocks 23 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 116%.
Scale-out only pays off if the host platform has enough PCIe lanes, slot spacing, power, and cooling.
The bigger the setup gets, the more the runtime matters. Multi-GPU and multi-user serving are where vLLM, SGLang, TGI, TensorRT-LLM, or tuned llama.cpp start to earn their complexity.
~$15,000 MSRP
NVIDIA H200 141GBNext step upUnlocks 2 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 20%.
~$30,000 MSRP
Unlocks 12 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 23%.
~$20,000 MSRP
Unlocks 13 additional models that do not fit on the current setup.
Lifts average decode speed across fitting models by about 38%.
~$8,000 MSRP
Gaudi 3 128GB (128 GB VRAM) can run these top models: Qwen 3.5 122B A10B (score: 99/100), Mistral Small 4 119B (score: 97/100), Devstral 2 123B Instruct (score: 97/100). See the full compatibility list above.
Gaudi 3 128GB has 128 GB of VRAM available for AI model inference. This determines which models and quantization levels you can run locally.
Yes, Gaudi 3 128GB is excellent for running LLMs locally with top compatibility scores above 80/100.
For coding on Gaudi 3 128GB, we recommend Qwen3-Coder-Next. It achieves 174.9 tokens per second with 256K context window. This model is a direct match for coding. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.
There are 4 upgrade path(s) from Gaudi 3 128GB: Gaudi 3 128GB, NVIDIA H200 141GB. Upgrading would unlock larger models and faster inference speeds.
Yes, Gaudi 3 128GB with 128 GB of usable memory can run Flux.1 Dev at FP16 natively. Flux is a 12B parameter diffusion transformer that produces high-quality images. You can also run the Schnell variant for faster generation.
Gaudi 3 128GB (128 GB VRAM) can handle various AI generation tasks beyond LLMs. For image generation, SDXL and Stable Diffusion 3.5 run well. Flux.1 Dev also runs natively for state-of-the-art image quality. For video, LTX Video 2.3 can generate short clips. Check the AI Capability Matrix above for detailed compatibility.
Gaudi 3 128GB is excellent for AI image generation. With 128 GB of usable memory, it runs all major diffusion models including Flux.1, SDXL, and Stable Diffusion 3.5 at full precision. You can generate high-resolution images quickly and even handle video generation models.
Yes, Gaudi 3 128GB with 128 GB of usable memory can run Qwen 3.5 27B at Q8 (near-lossless, ~28.9 GB) or even FP16 (~55.4 GB) depending on your context needs. This setup provides an excellent experience with this model. Use Ollama or vLLM for best results.
With 128 GB VRAM on Gaudi 3 128GB, use Q8_0 for most models — it is near-lossless and you have the memory for it. For 70B+ models, Q6_K offers excellent quality. Reserve Q4_K_M for 100B+ models or when you need maximum context length.
Gaudi 3 128GB already has strong memory bandwidth, so the next limit is often memory capacity and context headroom rather than raw decode speed. For local LLMs, fit first and bandwidth second is the right mental model.
Gaudi 3 128GB can be attractive on memory-per-dollar, but CUDA still has the broadest support across runtimes, kernels, guides, and community-tested local AI workflows. If your priority is the easiest setup and widest model compatibility, NVIDIA remains the safer choice. If your priority is value and you are comfortable with a narrower software stack, Gaudi 3 128GB can still be useful.
Gaudi 3 128GB supports up to 8× GPU scaling via PCIe. With 8× GPUs, you get 1024 GB effective memory with a 0.85× scaling factor per GPU. This enables running models like Qwen 3.5 397B A17B and Kimi K2.5 that don't fit on a single card.
Gaudi 3 128GB uses PCIe for multi-GPU communication, which has approximately 15% scaling overhead. For best multi-GPU performance, consider NVLink-equipped variants.
Usually yes. If you want to run 2-4× Gaudi 3 128GB for local AI, the bottleneck often becomes the platform, not the card. Workstation and server boards give you more CPU PCIe lanes, better x16 slot wiring, more spacing between cards, stronger power delivery, and usually more RAM capacity. Consumer x8/x8 layouts can work, but they are a common weak point in multi-GPU builds.
Compare with similar
