Info

Nothing is more valuable than intelligence. Luckily, inferencing, tuning and training gigantic cutting-edge LLMs has become a commodity. Thanks to state-of-the-art, open-source LLMs you can download for free, the only thing you need is suitabe hardware. We are proud to offer bleeding-edge Nvidia and AMD rack servers with the most competitive pricing in the world. From smaller air-cooled systems like the Nvidia GH200 Grace-Hopper Superchip and the HGX B200 and HGX B300 up to massive, NV-linked, liquid-cooled, CDU-integrated, ready-to-use Nvidia GB200 Grace-Blackwell Superchip and GB300 Grace-Blackwell Ultra Superchip systems. Multiple NV-linked (G)B200 or (G)B300 act as a single giant GPU with one single giant coherent memory pool. We are the only ones who offer smaller (half-size) systems than a complete NVL72 rack with "only" 18 superchips (NVL36). If you like AMD, we offer Mi325X and Mi350X systems too. All systems are perfect for inferencing insanely huge LLMs, quick fine-tuning and training of LLMs, image and video generation and editing as well as high-performance computing.

Example use case 1: Inferencing Deepseek R1 0528 685B, Nvidia Nemotron Ultra 253B, Llama 4 Maverick 400B, MiniMax-M1 456B, ERNIE-4.5-VL-424B-A47B, Kimi K2 1T, Qwen3 Coder 480B A35B or Qwen3-235B-A22B 2507
  • Deepseek R1 0528 685B: https://huggingface.co/deepseek-ai/DeepSeek-R1-0528
  • Nvidia Nemotron Ultra 253B: https://www.nvidia.com/en-us/ai-data-science/foundation-models/llama-nemotron/
  • Llama 4: https://llama.com
  • MiniMax-M1 456B: https://huggingface.co/MiniMaxAI/MiniMax-M1-80k
  • ERNIE-4.5-VL-424B-A47B: https://ernie.baidu.com/blog/posts/ernie4.5/
  • Kimi K2 1T: https://huggingface.co/moonshotai/Kimi-K2-Instruct
  • Qwen3 Coder 480B A35B: https://github.com/QwenLM/Qwen3-Coder
  • Qwen3-235B-A22B 2507: https://github.com/QwenLM/Qwen3/
    • Deepseek R1 0528 685B, Nvidia Nemotron Ultra 253B, Llama 4 Maverick 400B, MiniMax-M1 456B, ERNIE-4.5-VL-424B-A47B, Kimi K2 1T, Qwen3 Coder 480B A35B as well as Qwen3-235B-A22B 2507 are the most powerful open-source models by far and even match GPT-o1/o3/o4, Claude 4 Opus, Grok 4 and Gemini 2.5 Pro.
  • Deepseek R1 0528 685B with 4-bit quantization needs at least 410GB of memory to swiftly run inference! Nvidia Nemotron Ultra 253B with 8-bit quantization needs at least 270GB of memory to swiftly run inference! Qwen3-235B-A22B 2507 with 4-bit quantization needs at least 142GB of memory to swiftly run inference! Luckily, GH200 has a minimum of 576GB, GB300 a minimum of 784GB. With GH200 Qwen3-235B-A22B 2507 in 4bit can be run in VRAM only for ultra high inference speed (significantly more than 50 tokens/s). With Mi325/350 and GB200 Blackwell, as well as GB300 Blackwell Ultra this is also possible for Deepseek R1 0528 685B. With GB200 Blackwell and GB300 Blackwell Ultra you can expect up to 1000 tokens/s. If the model is bigger than VRAM you can only expect approx. 10-20 tokens/s. Surprisingly, Deepseek R1 0528 685B in 4-bit runs on our smallest system the GH200 with up to 20 tokens/s. That is usable! 4-bit quantization is the best trade-off between speed and accuracy, but is natively only supported by GB200 Blackwell and GB300 Blackwell Ultra. We recommend using Nvidia Dynamo (https://www.nvidia.com/en-us/ai/dynamo/) for inferencing.

    • Example use case 2: Fine-tuning Deepseek R1 0528 685B with PyTorch FSDP and Q-Lora
  • Tutorial: https://www.philschmid.de/fsdp-qlora-llama3
  • The ultimate guide to fine-tuning: https://arxiv.org/abs/2408.13296
  • Models need to be fine-tuned on your data to unlock the full potential of the model. But efficiently fine-tuning bigger models like Deepseek R1 0528 685B remained a challenge until now. This blog post walks you through how to fine-tune Deepseek R1 using PyTorch FSDP and Q-Lora with the help of Hugging Face TRL, Transformers, peft & datasets.
  • Fine-tuning big models within a reasonable time requires special and beefy hardware! Luckily, GH200, GB200/GB300 or Mi325/350 are ideal to complete this task extremely quickly.

    • Example use case 3: Generating videos with Mochi1, HunyuanVideo, MAGI-1 or Wan 2.1
  • Mochi1: https://github.com/genmoai/models
  • Tencent HunyuanVideo: https://aivideo.hunyuan.tencent.com/
  • MAGI-1: https://github.com/SandAI-org/Magi-1
  • Wan 2.1: https://github.com/Wan-Video/Wan2.1
  • Mochi1, HunyuanVideo, MAGI-1 and Wan 2.1 are democratizing efficient video production for all.
  • Generating videos with requires special and beefy hardware! Mochi1 and HunyuanVideo need at least 80GB of VRAM. Luckily, GH200, B300 and GB300 or Mi325/Mi350 are ideal for this task. GH200 has a minimum of 96GB, GB300 a minimum of 288GB, B200 a minimum of 1.5TB and Mi325X has a minimum of 2TB.

    • Example use case 4: Image generation with HiDream-I1, Flux.1 or SANA-Sprint.
  • HiDream-I1: https://github.com/HiDream-ai/HiDream-I1
  • Flux: https://github.com/black-forest-labs/flux
  • SANA-Sprint: https://nvlabs.github.io/Sana/Sprint/
  • HiDream-I1 and Flux.1 are the best image generators at the moment. And they are uncensored, too. SANA-Sprint is very fast and efficient.
  • In high-speed inference, FLUX requires approximately 33GB of VRAM for maximum speed. For training the FLUX model, more than 40GB of VRAM is needed. SANA-Sprint requires up to 67GB of VRAM. Luckily, GH200 has a minimum of 96GB, GB300 a minimum of 288GB, B200 a minimum of 1.5TB and Mi325X has a minimum of 2TB.

    • Example use case 5: Image editing with FLUX.1-Kontext-dev, Omnigen 2, Nvidia Add-it, HiDream-E1 or ICEdit.
  • FLUX.1-Kontext-dev: https://bfl.ai/announcements/flux-1-kontext-dev
  • Omnigen 2: https://github.com/VectorSpaceLab/OmniGen2
  • Nvidia Add-it: https://research.nvidia.com/labs/par/addit/
  • HiDream-E1: https://github.com/HiDream-ai/HiDream-E1
  • ICEdit: https://river-zhang.github.io/ICEdit-gh-pages/
  • Omnigen 2, Add-it, HiDream-E1 and ICEdit are the most innovative and easy to use image editors at the moment.
  • For maximum speed in high resolution image generation and editing beefier hardware than consumer graphics cards is needed. Luckily, GH200, GB200, GB300, Mi325 and Mi350 excel at this task.

    • Example use case 6: Video editing with AutoVFX, SkyReels-A2 or VACE
  • AutoVFX: https://haoyuhsu.github.io/autovfx-website/
  • SkyReels-A2: https://skyworkai.github.io/skyreels-a2.github.io/
  • VACE: https://ali-vilab.github.io/VACE-Page/
  • AutoVFX, SkyReels-A2 and VACE are the most innovative and easy to use video editor at the moment.
  • For maximum speed in high resolution video editing beefier hardware than consumer graphics cards is needed. Luckily, GH200, GB200, GB300, Mi325 and Mi350 excel at this task.

    • Example use case 7: Deep Research with WebThinker
  • WebThinker: https://github.com/RUC-NLPIR/WebThinker
  • WebThinker enables large reasoning models to autonomously search, deeply explore web pages, and draft research reports, all within their thinking process.
  • The hardware requirements for using Webthinker depend on the particular LLM of choice (see above).

    • Example use case 8: Creating a Large Language Model from scratch
  • Tutorial: https://www.pluralsight.com/resources/blog/data/how-build-large-language-model
  • Imagine stepping into the world of language models as a painter stepping in front of a blank canvas. The canvas here is the vast potential of Natural Language Processing (NLP), and your paintbrush is the understanding of Large Language Models (LLMs). This article aims to guide you, new to NLP, in creating your first Large Language Model from scratch, focusing on the Transformer architecture and utilizing TensorFlow and Keras.
  • Taining a LLM from scratch within a reasonable time requires special and extremely beefy hardware! Luckily, GH200, GB200 and GB300 or Mi325/Mi350 are ideal for this task.

    • Why should you buy your own hardware?
  • "You'll own nothing and you'll be happy?" No!!! Never should you bow to Satan and rent stuff that you can own. In other areas, renting stuff that you can own is very uncool and uncommon. Or would you prefer to rent "your" car instead of owning it? Most people prefer to own their car, because it's much cheaper, it's an asset that has value and it makes the owner proud and happy. The same is true for compute infrastructure.
  • Even more so, because data and compute infrastructure are of great value and importance and are preferably kept on premises, not only for privacy reasons but also to keep control and mitigate risks. If somebody else has your data and your compute infrastructure you are in big trouble.
  • Speed, latency and ease-of-use are also much better when you have direct physical access to your stuff.
  • With respect to AI and specifically LLMs there is another very important aspect. The first thing big tech taught their closed-source LLMs was to be "politically correct" (lie) and implement guardrails, "safety" and censorship to such an extent that the usefulness of these LLMs is severely limited. Luckily, the open-source tools are out there to build and tune AI that is really intelligent and really useful. But first, you need your own hardware to run it on.

    • What are the main benefits of GH200 Grace-Hopper, GB200 Grace-Blackwell and GB300 Grace-Blackwell Ultra?
  • They have enough memory to run, tune and train the biggest LLMs currently available.
  • Their performance in every regard is almost unreal (up to 10,000 times faster than x86).
  • There are no alternative systems with the same amount of memory.
  • Ideal for AI, especially inferencing, fine-tuning and training of LLMs.
  • Multiple NV-linked GB200 or GB300 act as a single giant GPU.
  • Optimized for memory-intensive AI and HPC performance.
  • Ideal for HPC applications like, e.g. vector databases.
  • Easily customizable, upgradable and repairable.
  • Privacy and independence from cloud providers.
  • Cheaper and much faster than cloud providers.
  • They can be very quiet (with liquid-liquid CDU).
  • Reliable and energy-efficient liquid cooling.
  • Flexibility and the possibility of offline use.
  • Gigantic amounts of coherent memory.
  • They are very power-efficient.
  • The lowest possible latency.
  • They are beautiful.
  • CUDA enabled.
  • Run Linux.

    • GB200 Blackwell and GB300 Blackwell Ultra

    The Nvidia GB200 Grace-Blackwell superchip has truly amazing specs to show off. GPTrack.ai ready-to-use rack server systems with multiple NV-linked Nvidia GB200 Grace-Blackwell (up to 72) are available now. GB300 Blackwell Ultra will be available in Q1 2026. Be one of the first in the world to get a GB200 or GB300 rack system. Order now!

    What is the difference to alternative systems?
    The main difference between GH200/GB200/GB300 and alternative systems is that with GH200/GB200/GB300, the GPU is connected to the CPU via a 900 GB/s chip-2-chip NVLink vs. 128 GB/s PCIe gen5 used by traditional systems. Furthermore, multiple GB200/GB300 superchips and HGX B200/B300 are connected via 1800 GB/s NVLink vs. orders of magnitude slower network or PCIe connections used by traditional systems. Since these are the main bottlenecks, GH200/GB200/GB300's high-speed connections directly translate to much higher performance compared to traditional architectures. Also, multiple NV-linked (G)B200 or (G)B300 act as a single giant GPU with one single giant coherent memory pool. Since even PCIe gen 6 is much slower, Nvidia does not offer B200 and B300 as PCIe cards any more (only as SXM or superchip). We highly recommend choosing NV-linked systems over systems connected via PCIe and/or network.

    What is the difference to server systems of competitors?
  • Pricing: We aim to offer the most competitive pricing worldwide.
  • Size: A single GB200/GB300 NVL72 rack gives you more than an exaflop of compute. For some people, one complete rack is more than needed and too expensive. That is why we also offer smaller (half-size) systems with only 18 superchips (NVL36). We are, to our knowledge, the only ones in the world where you can get systems smaller than a complete GB200/GB300 NVL72 rack.
  • In-rack CDU: Our rack server systems are available with liquid cooling and a CDU integrated directly into the rack. You can choose between an air-liquid and liquid-liquid CDU.
  • Ready-to-use: Our systems can be ordered fully integrated and ready-to-use. Everything that is needed is included and tested. All you need to do is plug your system in to run it.
  • Shipping: Free shipping worldwide.
  • Warranty: 3-year manufacturer's warranty.

    • Technical details of our GB200/GB300 rackserver systems
  • 21-inch OCP rack
  • Liquid-cooled
  • In-rack CDU (air-liquid or liquid-liquid)
  • Multiple Nvidia GB200 Grace-Blackwell Superchips
  • Multiple Nvidia GB300 Grace-Blackwell Ultra Superchips
  • Multiple 72-core Nvidia Grace CPUs
  • Multiple Nvidia Blackwell B200 Tensor Core GPUs
  • Multiple Nvidia Blackwell B300 Tensor Core GPUs
  • Up to 72x 480GB of LPDDR5X memory with error-correction code (ECC)
  • Up to 20TB of HBM3e memory
  • Up to 40TB of total fast-access memory
  • NVLink-C2C: 900 GB/s of bandwidth
  • NVLink-G2G: 1800 GB/s of bandwidth
  • Programmable from 1200W to 2800W TDP (CPU + 2 GPU + memory)
  • Total TDP per Rack up to 140 kW
  • Up to 8x power shelve
  • Up to 72x PCIe gen5 M.2 slots on board
  • Up to 288x PCIe gen5 drives (NVMe)
  • Up to 108x FHFL PCIe Gen5 x16
  • 3 years manufacturer's warranty
  • Up to 48U 600 x 2616 x 1200 mm (23.6 x 103 x 47.2")
  • Up to 1500 kg (3300 lbs)

    • Optional components
  • NIC Nvidia Bluefield-3
  • NIC Nvidia ConnectX-7/8
  • NIC Intel 100G
  • Up to 72x 4TB M.2 SSD
  • Up to 288x 8TB E1.S SSD
  • Up to 288x 60TB 2.5" SSD
  • Storage controller
  • Raid controller
  • OS preinstalled
  • Anything possible on request

    • Need something different? We are happy to build custom systems to your liking.

    Compute performance of one GH200 superchip
  • 67 teraFLOPS FP64
  • 1 petaFLOPS TF32
  • 2 petaFLOPS FP16
  • 4 petaFLOPS FP8

    • Compute performance of one GB200 superchip
  • 90 teraFLOPS FP64
  • 5 petaFLOPS TF32
  • 10 petaFLOPS FP16
  • 20 petaFLOPS FP8
  • 40 petaFLOPS FP4

    • Compute performance of one GB300 superchip
  • 3 teraFLOPS FP64
  • 5 petaFLOPS TF32
  • 10 petaFLOPS FP16
  • 20 petaFLOPS FP8
  • 40 petaFLOPS FP4

    • Benchmarks
  • https://github.com/mag-/gpu_benchmark
    • Phoronix has so far benchmarked the Grace CPU. More is coming soon:
  • https://www.phoronix.com/review/nvidia-gh200-gptshop-benchmark
  • https://www.phoronix.com/review/nvidia-gh200-amd-threadripper
  • https://www.phoronix.com/review/aarch64-64k-kernel-perf
  • https://www.phoronix.com/review/nvidia-gh200-compilers
  • https://www.phoronix.com/review/nvidia-grace-epyc-turin

    • Download

    Here you can find various downloads concerning our systems: operating systems, firmware, drivers, software, manuals, white papers, spec sheets and so on. Everything you need to run your system and more.

    Spec sheets
  • GH200 624GB: Spec sheet GH200 624GB.pdf
  • GH200 Giga 624GB: Spec sheet GH200 Giga 624GB.pdf
  • Mi325X Air 2TB: Spec sheet Mi325X Air 2TB.pdf
  • Mi325X Liquid 2TB: Spec sheet Mi325X Liquid 2TB.pdf
  • Mi350X Air 2.3TB: Spec sheet Mi350X Air 2.3TB.pdf
  • B200 Air 1.5TB: Spec sheet B200 Air 1.5TB.pdf
  • B200 Liquid 1.5TB: Spec sheet B200 Liquid 1.5TB.pdf
  • B300 Air 2.3TB: Spec sheet B300 Air 2.3TB.pdf
  • GB200 NVL36 15TB: Spec sheet GB200 NVL36 15TB.pdf
  • GB200 NVL72 30TB: Spec sheet GB200 NVL72 30TB.pdf
  • GB300 NVL36 20TB: Spec sheet GB300 NVL36 20TB.pdf
  • GB300 NVL72 40TB: Spec sheet GB300 NVL72 40TB.pdf

    • Manuals
  • Official Nvidia GH200 Manual: https://docs.nvidia.com/grace/#grace-hopper
  • Official Nvidia Grace Manual: https://docs.nvidia.com/grace/#grace-cpu
  • Official Nvidia Grace getting started: https://docs.nvidia.com/grace/#getting-started-with-nvidia-grace
  • GH200 624GB: Manual GH200 624GB.pdf
  • GH200 Giga 624GB: Manual GH200 Giga 624GB.pdf
  • GB200 NVL: Manual GB200 NVL.pdf

    • Operating systems for Nvidia systems
  • Ubuntu Server for ARM: https://cdimage.ubuntu.com/releases/24.04/release/ubuntu-24.04.2-live-server-arm64+largemem.iso

    • Using the newest Nvidia 64k kernel is highly recommended: https://packages.ubuntu.com/search?keywords=linux-nvidia-64k-hwe

    Operating systems for AMD systems
  • Ubuntu Server for x86: https://ubuntu.com/download/server

    • Any other x86 linux distribution with kernel > 6.8 should work just fine. Using the newest kernel is highly recommended.

    Drivers
  • Nvidia GH200 drivers: https://www.nvidia.com/Download/index.aspx?lang=en-us
    Select product type "data center", product series "HGX-Series" and operating system "Linux aarch64".
  • Aspeed drivers: https://aspeedtech.com/support_driver/
  • Nvidia Bluefield-3 drivers: https://developer.nvidia.com/networking/doca#downloads
  • Nvidia ConnectX drivers: https://network.nvidia.com/products/ethernet-drivers/linux/mlnx_en/
  • Intel E810-CQDA2 drivers: https://www.intel.com/content/www/us/en/download/19630/intel-network-adapter-driver-for-e810-series-devices-under-linux.html?wapkw=E810-CQDA2
  • Graid SupremeRAID drivers: https://docs.graidtech.com/#linux-driver

    • Firmware
  • GH200 BMC: GH200 BMC.zip
  • GH200 BIOS: GH200 BIOS.zip
  • Nvidia Bluefield-3 firmware: https://network.nvidia.com/support/firmware/bluefield3/
  • Nvidia ConnectX firmware: https://network.nvidia.com/support/firmware/firmware-downloads/
  • Intel E810-CQDA2 firmware: https://www.intel.com/content/www/us/en/search.html?ws=idsa-default#q=E810-CQDA2

    • White papers
  • Nvidia GH200 Grace-Hopper white paper
  • Nvidia Blackwell white paper
  • Developing for Nvidia superchips
  • The ultimate guide to fine-tuning
  • Diffusion LLMs
  • Absolute Zero
  • MatFormer
  • Puzzle - Inference-optimized LLMs
  • Self-Adapting Language Models
  • Apex
  • Mixture-of-Recursions
  • Flexolmo
  • Coconut

    • Top open-source LLMs
  • Deepseek R1 0528 685B: https://huggingface.co/deepseek-ai/DeepSeek-R1-0528
  • Qwen3 Coder 480B A35B: https://github.com/QwenLM/Qwen3-Coder
  • Qwen3-235B-A22B 2507: https://github.com/QwenLM/Qwen3
  • Nvidia Llama Nemotron Super 49B and Ultra 253B: https://www.nvidia.com/en-us/ai-data-science/foundation-models/llama-nemotron/
  • Dots.llm1 142B: https://github.com/rednote-hilab/dots.llm1
  • QwQ-32B: https://qwenlm.github.io/blog/qwq-32b/
  • Llama 3.1, 3.2, 3.3 and 4: https://www.llama.com/
  • Llama 4 Maverick 400B: https://huggingface.co/meta-llama/Llama-4-Maverick-17B-128E-Instruct
  • Llama 4 Scout 109B: https://huggingface.co/meta-llama/Llama-4-Scout-17B-16E-Instruct
  • Cogito v1 preview 70B: https://huggingface.co/deepcogito/cogito-v1-preview-qwen-32B
  • GLM-4-32B: https://github.com/THUDM/GLM-4
  • Mistral Large 2 123B: https://huggingface.co/mistralai/Mistral-Large-Instruct-2407
  • Pixtral Large 123B: https://mistral.ai/news/pixtral-large/
  • Llama-3.2 Vision 90B: https://huggingface.co/meta-llama/Llama-3.2-90B-Vision
  • Llama-3.1 405B: https://huggingface.co/meta-llama/Llama-3.1-405B
  • Deepseek V3 671B: https://huggingface.co/deepseek-ai/DeepSeek-V3
  • MiniMax-01 456B: https://www.minimaxi.com/en/news/minimax-01-series-2
  • Tülu 3 405B: https://allenai.org/tulu
  • Qwen2.5 VL 72B: https://huggingface.co/Qwen/Qwen2.5-VL-72B-Instruct
  • Aya Vision: https://cohere.com/blog/aya-vision
  • Gemma-3 27B: https://blog.google/technology/developers/gemma-3/
  • Mistral Small 3.1 24B: https://mistral.ai/news/mistral-small-3-1
  • EXAONE Deep 32B: https://github.com/LG-AI-EXAONE/EXAONE-Deep
  • Skywork-R1V 38B: https://github.com/SkyworkAI/Skywork-R1V
  • Phi-4 models: https://azure.microsoft.com/en-us/blog/one-year-of-phi-small-language-models-making-big-leaps-in-ai/
  • Seed1.5-VL: https://seed.bytedance.com/en/tech/seed1_5_vl
  • BAGEL 14B: https://bagel-ai.org/
  • UniVG-R1 8B: https://amap-ml.github.io/UniVG-R1-page/
  • Magistral-Small-2506 24B: https://huggingface.co/mistralai/Magistral-Small-2506
  • MiniMax-M1: https://github.com/MiniMax-AI/MiniMax-M1
  • Tencent Hunyuan-A52B: https://github.com/Tencent-Hunyuan/Tencent-Hunyuan-Large
  • Tencent Hunyuan-A13B: https://github.com/Tencent-Hunyuan/Hunyuan-A13B
  • Baidu ERNIE 4.5: https://huggingface.co/collections/baidu/ernie-45-6861cd4c9be84540645f35c9
  • Ai2 OLMo 2: https://allenai.org/olmo
  • Pangu-Pro-MOE-72B: https://huggingface.co/IntervitensInc/pangu-pro-moe-model
  • Kimi K2 1T: https://huggingface.co/moonshotai/Kimi-K2-Instruct

    • Software
  • Nvidia Dynamo: https://www.nvidia.com/en-us/ai/dynamo/
  • Nvidia Github: https://github.com/NVIDIA
  • Nvidia CUDA: https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=arm64-sbsa
  • Nvidia Container-toolkit: https://github.com/NVIDIA/nvidia-container-toolkit
  • Nvidia Tensorflow: https://github.com/NVIDIA/tensorflow
  • Nvidia Pytorch: https://catalog.ngc.nvidia.com/orgs/nvidia/containers/pytorch
  • AMD ROCm: https://www.amd.com/en/products/software/rocm.html
  • Keras: https://keras.io/
  • Apache OpenNLP: https://opennlp.apache.org/
  • Nvidia NIM models: https://build.nvidia.com/explore/discover
  • Nvidia Triton inference server: https://www.nvidia.com/en-us/ai-data-science/products/triton-inference-server/
  • Nvidia NeMo Customizer: https://developer.nvidia.com/blog/fine-tune-and-align-llms-easily-with-nvidia-nemo-customizer/
  • Huggingface open-source LLMs: https://huggingface.co/models
  • Huggingface text generation inference: https://github.com/huggingface/text-generation-inference
  • vLLM - inference and serving engine: https://github.com/vllm-project/vllm
  • vLLM docker image: https://hub.docker.com/r/drikster80/vllm-gh200-openai
  • Ollama - run LLMs locally: https://ollama.com/
  • Llama.cpp: https://github.com/ggml-org/llama.cpp
  • Open WebUI: https://openwebui.com/
  • ComfyUI: https://www.comfy.org/
  • LM Studio: https://lmstudio.ai/
  • Llamafile: https://github.com/Mozilla-Ocho/llamafile
  • Fine-tune Llama 3 with PyTorch FSDP and Q-Lora: https://www.philschmid.de/fsdp-qlora-llama3/
  • Perplexica: https://github.com/ItzCrazyKns/Perplexica
  • Morphic: https://github.com/miurla/morphic
  • Open-Sora: https://github.com/hpcaitech/Open-Sora
  • Flux.1: https://github.com/black-forest-labs/flux
  • Storm: https://github.com/stanford-oval/storm
  • Stable Diffusion 3.5: https://huggingface.co/stabilityai/stable-diffusion-3.5-large
  • Genmo Mochi1: https://github.com/genmoai/models
  • Genmo Mochi1 (reduced VRAM): https://github.com/victorchall/genmoai-smol
  • Rhymes AI Allegro: https://github.com/rhymes-ai/Allegro
  • OmniGen: https://github.com/VectorSpaceLab/OmniGen
  • Segment anything: https://github.com/facebookresearch/segment-anything
  • AutoVFX: https://haoyuhsu.github.io/autovfx-website/
  • DimensionX: https://chenshuo20.github.io/DimensionX/
  • Nvidia Add-it: https://research.nvidia.com/labs/par/addit/
  • MagicQuill: https://magicquill.art/demo/
  • AnythingLLM: https://github.com/Mintplex-Labs/anything-llm
  • Pyramid-Flow: https://pyramid-flow.github.io/
  • LTX-Video: https://github.com/Lightricks/LTX-Video
  • CogVideoX: https://github.com/THUDM/CogVideo
  • OmniControl: https://github.com/Yuanshi9815/OminiControl
  • Samurai: https://yangchris11.github.io/samurai/
  • All Hands: https://www.all-hands.dev/
  • Tencent HunyuanVideo: https://aivideo.hunyuan.tencent.com/
  • Aider: https://aider.chat/
  • Unsloth: https://github.com/unslothai/unsloth
  • Axolotl: https://github.com/axolotl-ai-cloud/axolotl
  • Star: https://nju-pcalab.github.io/projects/STAR/
  • Sana: https://nvlabs.github.io/Sana/
  • RepVideo: https://vchitect.github.io/RepVid-Webpage/
  • UI-TARS: https://github.com/bytedance/UI-TARS
  • DiffuEraser: https://lixiaowen-xw.github.io/DiffuEraser-page/
  • Go-with-the-Flow: https://eyeline-research.github.io/Go-with-the-Flow/
  • 3DTrajMaster: https://fuxiao0719.github.io/projects/3dtrajmaster/
  • YuE: https://map-yue.github.io/
  • DynVFX: https://dynvfx.github.io/
  • ReasonerAgent: https://reasoner-agent.maitrix.org/
  • Open-source DeepResearch: https://huggingface.co/blog/open-deep-research
  • Deepscaler: https://github.com/agentica-project/deepscaler
  • InspireMusic: https://funaudiollm.github.io/inspiremusic/
  • FlashVideo: https://github.com/FoundationVision/FlashVideo
  • MatAnyone: https://pq-yang.github.io/projects/MatAnyone/
  • LocalAI: https://localai.io/
  • Stepvideo: https://huggingface.co/stepfun-ai/stepvideo-t2v
  • SkyReels: https://github.com/SkyworkAI/SkyReels-V2
  • OctoTools: https://octotools.github.io/
  • SynCD: https://www.cs.cmu.edu/~syncd-project/
  • Mobius: https://mobius-diffusion.github.io/
  • Wan 2.1: https://github.com/Wan-Video/Wan2.1
  • TheoremExplainAgent: https://tiger-ai-lab.github.io/TheoremExplainAgent/
  • RIFLEx: https://riflex-video.github.io/
  • Browser use: https://browser-use.com/
  • HunyuanVideo-I2V: https://github.com/Tencent/HunyuanVideo-I2V
  • Spark-TTS: https://sparkaudio.github.io/spark-tts/
  • GEN3C: https://research.nvidia.com/labs/toronto-ai/GEN3C/
  • DiffRhythm: https://aslp-lab.github.io/DiffRhythm.github.io/
  • Babel: https://babel-llm.github.io/babel-llm/
  • Diffusion Self-Distillation: https://primecai.github.io/dsd/
  • OWL: https://github.com/camel-ai/owl
  • ANUS: https://github.com/nikmcfly/ANUS
  • Long Context Tuning for Video Generation: https://guoyww.github.io/projects/long-context-video/
  • Tight Inversion: https://tight-inversion.github.io/
  • VACE: https://ali-vilab.github.io/VACE-Page/
  • SANA-Sprint: https://nvlabs.github.io/Sana/Sprint/
  • Sesame Conversational Speech Model: https://github.com/SesameAILabs/csm
  • Search-R1: https://github.com/PeterGriffinJin/Search-R1
  • AI Scientist: https://github.com/SakanaAI/AI-Scientist
  • SpatialLM: https://manycore-research.github.io/SpatialLM/
  • Nvidia Cosmos: https://www.nvidia.com/en-us/ai/cosmos/
  • AudioX: https://zeyuet.github.io/AudioX/
  • AccVideo: https://aejion.github.io/accvideo/
  • Video-T1: https://liuff19.github.io/Video-T1/
  • InfiniteYou: https://bytedance.github.io/InfiniteYou/
  • BizGen: https://bizgen-msra.github.io/
  • ParetoQ: https://github.com/facebookresearch/ParetoQ
  • DAPO: https://dapo-sia.github.io/
  • OpenDeepSearch: https://github.com/sentient-agi/OpenDeepSearch
  • KTransformers: https://github.com/kvcache-ai/ktransformers
  • SkyReels-A2: https://skyworkai.github.io/skyreels-a2.github.io/
  • Human Skeleton and Mesh Recovery: https://isshikihugh.github.io/HSMR/
  • Segment Any Motion in Videos: https://motion-seg.github.io/
  • Lumina-mGPT 2.0: https://github.com/Alpha-VLLM/Lumina-mGPT-2.0
  • HiDream-I1: https://github.com/HiDream-ai/HiDream-I1
  • HiDream-E1: https://github.com/HiDream-ai/HiDream-E1
  • Transformer Lab: https://transformerlab.ai/
  • One-Minute Video Generationwith Test-Time Training: https://test-time-training.github.io/video-dit/
  • OmniTalker: https://humanaigc.github.io/omnitalker/
  • UNO: https://bytedance.github.io/UNO/
  • Skywork-OR1: https://github.com/SkyworkAI/Skywork-OR1
  • LettuceDetect: https://github.com/KRLabsOrg/LettuceDetect
  • Sonic: https://jixiaozhong.github.io/Sonic/
  • InstantCharacter: https://instantcharacter.github.io/
  • Dive: https://github.com/OpenAgentPlatform/Dive
  • Nari Dia-1.6B: https://github.com/nari-labs/dia
  • FramePack: https://github.com/lllyasviel/FramePack
  • RAGEN: https://github.com/RAGEN-AI/RAGEN
  • LiveCC: https://showlab.github.io/livecc/
  • Reflection2perfection: https://diffusion-cot.github.io/reflection2perfection/
  • UNI3C: https://ewrfcas.github.io/Uni3C/
  • MAGI-1: https://github.com/SandAI-org/Magi-1
  • Parakeet TDT 0.6B V2: https://huggingface.co/nvidia/parakeet-tdt-0.6b-v2
  • VRAM Calculator: https://apxml.com/tools/vram-calculator
  • ICEdit: https://river-zhang.github.io/ICEdit-gh-pages/
  • FantasyTalking: https://fantasy-amap.github.io/fantasy-talking/
  • 3DV-TON: https://2y7c3.github.io/3DV-TON/
  • WebThinker: https://github.com/RUC-NLPIR/WebThinker
  • SGLang: https://github.com/sgl-project/sglang
  • Jan: https://jan.ai/
  • Rasa: https://rasa.com/
  • ACE-Step: https://ace-step.github.io/
  • AIQToolkit: https://github.com/NVIDIA/AIQToolkit
  • ZeroSearch: https://alibaba-nlp.github.io/ZeroSearch/
  • DreamO: https://github.com/bytedance/DreamO
  • HoloTime: https://zhouhyocean.github.io/holotime/
  • FlexiAct: https://shiyi-zh0408.github.io/projectpages/FlexiAct/
  • HunyuanCustom: https://hunyuancustom.github.io/
  • PixelHacker: https://hustvl.github.io/PixelHacker/
  • ZenCtrl: https://fotographer.ai/zenctrl
  • T2I-R1: https://github.com/CaraJ7/T2I-R1
  • TFrameX: https://github.com/TesslateAI/TFrameX
  • Tesslate Studio: https://github.com/TesslateAI/Studio
  • Continuous Thought Machines: https://github.com/SakanaAI/continuous-thought-machines
  • NVIDIA Isaac GR00T: https://github.com/NVIDIA/Isaac-GR00T
  • BLIP3-o: https://github.com/JiuhaiChen/BLIP3o
  • DeerFlow: https://github.com/bytedance/deer-flow
  • Clara: https://github.com/badboysm890/ClaraVerse
  • Void: https://voideditor.com/
  • LLM-d: https://github.com/llm-d/llm-d
  • Roocode: https://roocode.com/
  • MCP Filesystem Server: https://github.com/mark3labs/mcp-filesystem-server
  • Surfsense: https://github.com/MODSetter/SurfSense
  • S3: https://github.com/pat-jj/s3
  • Ramalama: https://ramalama.ai/
  • FLUX.1 Kontext: https://bfl.ai/models/flux-kontext
  • ExLlamaV2: https://github.com/turboderp-org/exllamav2
  • MLC LLM: https://github.com/mlc-ai/mlc-llm
  • LMDeploy: https://github.com/InternLM/lmdeploy
  • HunyuanVideo-Avatar: https://hunyuanvideo-avatar.github.io/
  • OmniConsistency: https://github.com/showlab/OmniConsistency
  • Phantom: https://github.com/Phantom-video/Phantom
  • Chatterbox TTS: https://github.com/resemble-ai/chatterbox
  • Gemini Fullstack LangGraph: https://github.com/google-gemini/gemini-fullstack-langgraph-quickstart
  • Ragbits: https://github.com/deepsense-ai/ragbits
  • Sparse transformers: https://github.com/NimbleEdge/sparse_transformers
  • Tokasaurus: https://github.com/ScalingIntelligence/tokasaurus
  • DeepVerse: https://sotamak1r.github.io/deepverse/
  • SkyReels-Audio: https://skyworkai.github.io/skyreels-audio.github.io/
  • HunyuanCustom: https://hunyuancustom.github.io/
  • OpenAudio: https://github.com/fishaudio
  • KVzip: https://github.com/snu-mllab/KVzip
  • Hcompany: https://www.hcompany.ai/
  • Text-to-LoRA: https://github.com/SakanaAI/text-to-lora
  • Rvn-tools: https://github.com/rvnllm/rvn-tools
  • Transformer Lab: https://github.com/transformerlab/transformerlab-app
  • Llama.cpp Server Launcher: https://github.com/thad0ctor/llama-server-launcher
  • PlayerOne: https://playerone-hku.github.io/
  • SeedVR2: https://iceclear.github.io/projects/seedvr2/
  • Any-to-Bokeh: https://vivocameraresearch.github.io/any2bokeh/
  • LayerFlow: https://sihuiji.github.io/LayerFlow-Page/
  • ImmerseGen: https://immersegen.github.io/
  • InterActHuman: https://zhenzhiwang.github.io/interacthuman/
  • LoRA-Edit: https://cjeen.github.io/LoraEditPaper/
  • LMCache: https://github.com/LMCache/LMCache
  • Polaris: https://github.com/ChenxinAn-fdu/POLARIS
  • LLM Visualization: https://bbycroft.net/llm
  • OmniGen2: https://github.com/VectorSpaceLab/OmniGen2
  • FLUX.1 Kontext [dev]: https://huggingface.co/black-forest-labs/FLUX.1-Kontext-dev
  • VMem: https://v-mem.github.io/
  • SongBloom: https://github.com/Cypress-Yang/SongBloom
  • Hunyuan-GameCraft: https://hunyuan-gamecraft.github.io/
  • Unmute: https://github.com/kyutai-labs/unmute
  • EX-4D: https://tau-yihouxiang.github.io/projects/EX-4D/EX-4D.html
  • XVerse: https://github.com/bytedance/XVerse
  • MemOS: https://memos.openmem.net/
  • Opencode: https://github.com/sst/opencode
  • StreamDiT: https://cumulo-autumn.github.io/StreamDiT/
  • ThinkSound: https://thinksound-project.github.io/
  • OmniVCus: https://caiyuanhao1998.github.io/project/OmniVCus/
  • Universal Tool Calling Protocol (UTCP): https://github.com/universal-tool-calling-protocol
  • Agent Reinforcement Trainer: https://github.com/OpenPipe/ART
  • MeiGen MultiTalk: https://meigen-ai.github.io/multi-talk/
  • ik_llama.cpp: https://github.com/ikawrakow/ik_llama.cpp
  • NeuralOS: https://neural-os.com/
  • System prompts: https://github.com/x1xhlol/system-prompts-and-models-of-ai-tools
  • GPUstack: https://github.com/gpustack/gpustack
  • Rust GPU: https://rust-gpu.github.io/

    • Benchmarking
  • GPU benchmark: https://github.com/mag-/gpu_benchmark
  • Ollama benchmark: https://llm.aidatatools.com/results-linux.php
  • Phoronix test suite: https://www.phoronix-test-suite.com/
  • MLCommons: https://mlcommons.org/benchmarks/
  • Artifical Analysis: https://artificialanalysis.ai/
  • Lmarena: https://lmarena.ai/
  • Livebench: https://livebench.ai/
  • AMD vs NVIDIA Inference Benchmark: https://semianalysis.com/2025/05/23/amd-vs-nvidia-inference-benchmark-who-wins-performance-cost-per-million-tokens/
  • Openrouter Rankings: https://openrouter.ai/rankings
  • NeuSight: https://github.com/sitar-lab/NeuSight

    • Contact

    Email: x@GPTrack.ai

    GPT LLC
    Fifth Floor, Zephyr House, 122 Mary Street
    George Town, P.O. Box 31493
    Grand Cayman KY1-1206
    Cayman Islands
    Company register number: HM-7509

    Free shipping worldwide. 3-year manufacturer's warranty. Our prices do not contain any taxes. We accept almost all currencies there are. Payment is possible via wire transfer or cash.

    Try

    Try before you buy. You can apply for remote testing. You will then be given login credentials for remote access. If you want to come by and see for yourself and run some tests, that is also possible any time.

    Currently available for testing:

  • GH200 624GB
  • GH200 Giga 624GB
  • GH200 2x576GB
  • Mi325X 2TB
  • GB200 NVL72 30TB

    • Apply via email: x@GPTrack.ai