DGX H100 · 8 × 80 GB
NVIDIA · DGX H100
207 text models run well · 3,350 GB/s bandwidth
Models on this hardware
Start here
File UD-Q4_K_M · 16.9 GB · unsloth/gemma-4-26B-A4B-it-GGUF · What is Q4?
Memory Needs ~19 GB · ~35 GB on DGX H100 · 640 GB VRAM, including ~16 GB the system keeps
ollama run hf.co/unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
File Q4_K_M · 16.5 GB · prithivMLmods/Smaug-Mini-GGUF · What is Q4?
Memory Needs ~21 GB · ~37 GB on DGX H100 · 640 GB VRAM, including ~16 GB the system keeps
ollama run hf.co/prithivMLmods/Smaug-Mini-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
Text
Text · LiveBench · overall
0–100 · higher is better
- Color = publisher
Source: LiveBench 2026-06-25 · collected 2026-09-26 · Open models without a score are not drawn (no score ≠ 0).
| Smaug MiniQ4 · published params · source | 76.9 | Not on the LMArena text leaderboard of 2026-10-22 | ~37 GB | Q4 · published params |
|---|---|---|---|---|
| GLM-5.3Q4 · published params · source | 76.1 | Arena lists only GLM-5.3 (Max), a named effort setting, so no rating is matched to this entry | ~574 GB | Q4 · published params |
| Qwen3.8 27BQ4 · published params · source | 75.3 | 1,437 ± 5 | ~37 GB | Q4 · published params |
| DeepSeek V4 Flash · 0731Q4 · published params · source | 74.2 | Not on the LMArena text leaderboard of 2026-10-22 | ~244 GB | Q4 · published params |
| GLM-5.2Q4 · published params · source | 73.2 | Arena lists only GLM-5.2 (Max), a named effort setting, so no rating is matched to this entry | ~574 GB | Q4 · published params |
| GLM-5.3 FlashQ4 · published params · source | 71.6 | 1,476 ± 5 | ~256 GB | Q4 · published params |
| Nemotron 3 Ultra 550B A55BQ4 · published params · source | 67.4 | 1,426 ± 7 | ~429 GB | Q4 · published params |
| MiniMax M3Q4 · published params · source | 67.3 | 1,440 ± 4 | ~337 GB | Q4 · published params |
LMArena ratings in this table: LMArena 2026-10-22 · collected 2026-10-25 · dataset · CC BY 4.0
Every reviewed text model here 47
Every one that runs here, well or slowly, scored or not; 27 of 47 have a LiveBench or LMArena result. Run well first, then tight fits, then slow; within each, by Hub downloads over 30 days, read 2026-10-25, not by quality.
| Qwen3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~28 GB | 9.0 | Apache 2.0 | 10.4M |
|---|---|---|---|---|---|---|
| Gemma 4 31B | Not in LiveBench 2026-06-25 | 1,452 | ~40 GB | 18.3 | Apache 2.0 | 8.4M |
| Gemma 4 26B A4B | Not in LiveBench 2026-06-25 | 1,437 | ~35 GB | 16.9 | Apache 2.0 | 8.2M |
| Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GB | 5.0 | Apache 2.0 | 8.1M |
| Qwen3.8 27B | 75.3 | 1,437 | ~37 GB | 16.5 | Apache 2.0 | 7.8M |
| GLM-5.3 Flash | 71.6 | 1,476 | ~256 GB | 156.8 | MIT | 6.4M |
| Qwen3.5 9B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GB | 5.7 | Apache 2.0 | 6.1M |
| gpt-oss-20b | Not in LiveBench 2026-06-25 | 1,317 | ~32 GB | 11.6 | Apache 2.0 | 5.8M |
| Llama 3.1 8B Instruct | Not in LiveBench 2026-06-25 | 1,211 | ~22 GB | 4.9 | Llama 3.1 | 4.8M |
| DeepSeek V4 Flash · 0731 | 74.2 | Not on the LMArena text leaderboard of 2026-10-22 | ~244 GB | 136.7 | MIT | 4.1M |
| Qwen3 1.7B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~18 GB | 1.1 | Apache 2.0 | 4.1M |
| Gemma 4 E4B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~22 GB | 5.0 | Apache 2.0 | 4M |
| Llama 3.2 1B Instruct | Not in LiveBench 2026-06-25 | 1,111 | ~17 GB | 0.8 | Llama 3.2 | 3.5M |
| Gemma 4 E2B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~20 GB | 3.1 | Apache 2.0 | 3M |
| Qwen3 4B Instruct 2507 | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~19 GB | 2.5 | Apache 2.0 | 2.6M |
| Qwen3.6 27B | 64.0 | Not on the LMArena text leaderboard of 2026-10-22 | ~37 GB | 16.8 | Apache 2.0 | 2.4M |
| GLM-5.3 | 76.1 | Arena lists only GLM-5.3 (Max), a named effort setting, so no rating is matched to this entry | ~574 GB | 365.3 | GLM-5.3 | 2.1M |
| Gemma 4 12B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~25 GB | 7.1 | Apache 2.0 | 1.7M |
| Qwen3.5 35B A3B | Not in LiveBench 2026-06-25 | 1,394 | ~43 GB | 22.0 | Apache 2.0 | 1.5M |
| Nemotron 3 Nano 30B A3B | Not in LiveBench 2026-06-25 | 1,314 | ~46 GB | 24.6 | NVIDIA Nemotron Open Model | 1.4M |
| Llama 3.2 3B Instruct | Not in LiveBench 2026-06-25 | 1,167 | ~19 GB | 2.0 | Llama 3.2 | 1.1M |
| Qwen3 30B A3B Instruct 2507 | Not in LiveBench 2026-06-25 | 1,382 | ~39 GB | 18.6 | Apache 2.0 | 1.1M |
| GLM-4.7 Flash | Not in LiveBench 2026-06-25 | 1,365 | ~39 GB | 18.3 | MIT | 953.7K |
| DeepSeek V4 Flash | 65.5 | 1,436 | ~229 GB | 137.9 | MIT | 936.1K |
| Nemotron 3.5 Lightning 30B A3B | Not in LiveBench 2026-06-25 | Arena rated the NVFP4 build, NVIDIA-Nemotron-3.5-Lightning-30B-A3B (NVFP4); this entry is the BF16 reference, which NVIDIA reports with different scores | ~39 GB | 18.9 | OpenMDW 1.1 | 861.5K |
| Qwen3 Coder 30B A3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~39 GB | 18.6 | Apache 2.0 | 745K |
| SmolLM3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~19 GB | 1.9 | Apache 2.0 | 718.7K |
| Nemotron 3 Super 120B A12B | Not in LiveBench 2026-06-25 | Arena does not state the precision of the build it served, so no rating is matched to this NVFP4 checkpoint | ~106 GB | 82.5 | NVIDIA Nemotron Open Model | 535.1K |
| Phi-4 | Not in LiveBench 2026-06-25 | 1,256 | ~27 GB | 8.0 | MIT | 368.3K |
| DeepSeek R1 0528 Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GB | 5.0 | MIT | 355.8K |
| Phi-4 mini | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~19 GB | 2.5 | MIT | 342.8K |
| GLM-5.2 | 73.2 | Arena lists only GLM-5.2 (Max), a named effort setting, so no rating is matched to this entry | ~574 GB | 465.8 | MIT | 298.6K |
| Olmo 3 7B Instruct | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~22 GB | 4.5 | Apache 2.0 | 262.5K |
| MiniMax M3 | 67.3 | 1,440 | ~337 GB | 264.9 | MiniMax Community | 201.1K |
| Devstral Small 2 24B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~34 GB | 14.3 | Apache 2.0 | 195.6K |
| Mistral Small 3.2 24B | Not in LiveBench 2026-06-25 | 1,356 | ~35 GB | 14.3 | Apache 2.0 | 181.4K |
| Ministral 3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~27 GB | 8.2 | Apache 2.0 | 126.5K |
| MiMo V2.6 Flash | Not in LiveBench 2026-06-25 | 1,448 | ~248 GB | 126.4 | MIT | 89.6K |
| Ministral 3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~19 GB | 2.1 | Apache 2.0 | 87.4K |
| Ministral 3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GB | 5.2 | Apache 2.0 | 79.4K |
| Granite 4.2 8B | Not in LiveBench 2026-06-25 | 1,290 | ~22 GB | 5.3 | Apache 2.0 | 78.3K |
| Nemotron 3 Ultra 550B A55B | 67.4 | 1,426 | ~429 GB | 359.2 | OpenMDW 1.1 | 43.7K |
| Granite 4.2 3B | Not in LiveBench 2026-06-25 | 1,290 | ~19 GB | 2.2 | Apache 2.0 | 39.8K |
| Granite 4.2 30B | Not in LiveBench 2026-06-25 | 1,340 | ~39 GB | 17.7 | Apache 2.0 | 39.3K |
| Olmo 3 32B Think | Not in LiveBench 2026-06-25 | 1,306 | ~40 GB | 19.5 | Apache 2.0 | 13.1K |
| Ling 3.0 Flash | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~109 GB | 77.8 | MIT | 9.7K |
| Smaug Mini | 76.9 | Not on the LMArena text leaderboard of 2026-10-22 | ~37 GB | 16.5 | Apache 2.0 | 817 |
Adjust the estimate: quantization, context, reserve
Vision
Vision · LMArena
rating · higher is better
- Color = publisher
Source: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0 · Models without a rating are left out, not scored as zero.
| GLM-5.3 FlashQ4 · published params · source | 1,271 ± 6 | ~256 GB | Q4 · published params |
|---|---|---|---|
| Gemma 4 31BQ4 · published params · source | 1,262 ± 6 | ~40 GB | Q4 · published params |
| Qwen3.8 27BQ4 · published params · source | 1,242 ± 8 | ~37 GB | Q4 · published params |
| Gemma 4 26B A4BQ4 · published params · source | 1,242 ± 6 | ~35 GB | Q4 · published params |
| MiniMax M3Q4 · published params · source | 1,233 ± 6 | ~337 GB | Q4 · published params |
| Qwen3-VL-235B-A22B-InstructQ4 · Hugging Face count · source | 1,215 ± 7 | ~193 GB | Q4 · Hugging Face count |
| gemma-3-27b-itQ4 · Hugging Face count · source | 1,160 ± 8 | ~37 GB | Q4 · Hugging Face count |
| GLM-4.5VQ4 · Hugging Face count · source | 1,153 ± 12 | ~97 GB | Q4 · Hugging Face count |
LMArena ratings in this table: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0
12 more vision models in the catalog have no usable memory figure or parameter count for this profile. Browse Vision models
Image
Image · Prompt alignment · GenEval
0–1 · higher is better
Source: GenEval · 2023 · reviewed 2026-10-05 · Published quality in the original benchmark preset; not measured on this hardware.
| Stable Diffusion XL 1.0Platform path documented · stable-diffusion.cpp · source | 0.55 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|---|
| Stable Diffusion 1.5Platform path documented · stable-diffusion.cpp · source | 0.43 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
| Cosmos3-Super-Text2ImagePublisher minimum · CUDA · source | No comparable published result | 1,152 ± 6 | ~640 GB | Publisher minimum |
| Cosmos3-Super-Text2Image (agentic)Publisher minimum · CUDA · source | No comparable published result | 1,165 ± 7 | ~640 GB | Publisher minimum |
| Cosmos3-Super-Text2Image-4StepPublisher minimum · CUDA · source | No comparable published result | Arena rated Cosmos3-Super-Text2Image; this entry is the 4-step distilled checkpoint | ~640 GB | Publisher minimum |
| HunyuanImage 3.0 Instruct (Fal)Publisher minimum · CUDA · source | No comparable published result | Arena rated hunyuan-image-3.0; this entry is the Instruct checkpoint | ~640 GB | Publisher minimum |
| HunyuanImage 3.0 (Fal)Publisher minimum · CUDA · source | No comparable published result | 1,150 ± 3 | ~240 GB | Publisher minimum |
| Emu3.5-Image16-bit est. · params · source | No comparable published result | Not reviewed for LMArena | ~84 GB | 16-bit est. · params |
LMArena ratings in this table: LMArena 2026-10-07 · collected 2026-10-25 · dataset · CC BY 4.0
17 more image models in the catalog have no usable memory figure or parameter count for this profile. Browse Image models
Video
| Cosmos3-Super-Image2Video16-bit est. · params · source | ~158 GB | 16-bit est. · params |
|---|---|---|
| Cosmos3-Super-Image2Video-4Step16-bit est. · params · source | ~156 GB | 16-bit est. · params |
| MiniMax H316-bit est. · params · source | ~82 GB | 16-bit est. · params |
| Cosmos-H-SurgicalConditional path · CUDA · source | ~80 GB | Conditional path |
| Wan-Dancer-14BConditional path · CUDA · source | ~80 GB | Conditional path |
| Wan2.2 T2V-A14BConditional path · CUDA · source | ~80 GB | Conditional path |
| Wan2.2-I2V-A14BConditional path · CUDA · source | ~80 GB | Conditional path |
| Wan2.2-I2V-A14B-DiffusersConditional path · CUDA · source | ~80 GB | Conditional path |
How this list works Documented execution paths on this hardware first, then publisher minimums (CUDA or Apple MPS).
20 more video models in the catalog have no usable memory figure or parameter count for this profile. Browse Video models
Audio
Audio · Word error rate · Open ASR
% · lower is better
Source: Word error rate · Open ASR · reviewed 2026-10-20 · Published quality in the original benchmark preset; not measured on this hardware.
| Whisper large-v3-turboPlatform path documented · whisper.cpp / CUDA · source | 7.83 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|
| Whisper large-v3Platform path documented · whisper.cpp / CUDA · source | 7.44 | Documented path without a memory figure | Platform path documented |
| Step-Audio-R1.116-bit est. · params · source | No comparable published result | ~83 GB | 16-bit est. · params |
| Voxtral-Small-24B-250716-bit est. · params · source | No comparable published result | ~61 GB | 16-bit est. · params |
| Kimi-Audio-7B16-bit est. · params · source | No comparable published result | ~26 GB | 16-bit est. · params |
| Kimi-Audio-7B-Instruct16-bit est. · params · source | No comparable published result | ~26 GB | 16-bit est. · params |
| AuKPublisher minimum · CUDA · source | No comparable published result | ~25 GB | Publisher minimum |
| AuK-FlashPublisher minimum · CUDA · source | No comparable published result | ~25 GB | Publisher minimum |
33 more audio models in the catalog have no usable memory figure or parameter count for this profile. Browse Audio models
Embeddings
Embeddings · MTEB (Multilingual, v2) · mean(task)
0–100 · higher is better
- Color = publisher
| Qwen3-Embedding 8B8-bit · published params · source | 70.6 | ~12 GB | 8-bit · published params |
|---|---|---|---|
| Qwen3-Embedding 4B8-bit · published params · source | 69.5 | ~7 GB | 8-bit · published params |
| Harrier OSS v1 (0.6B)8-bit · published params · source | 69.0 | ~3 GB | 8-bit · published params |
| Qwen3-Embedding 0.6B8-bit · published params · source | 64.3 | ~3 GB | 8-bit · published params |
| Multilingual E5 Large Instruct8-bit · published params · source | 63.2 | ~3 GB | 8-bit · published params |
| EmbeddingGemma 28-bit · published params · source | 61.4 | ~3 GB | 8-bit · published params |
| EmbeddingGemma (300M)8-bit · published params · source | 61.1 | ~3 GB | 8-bit · published params |
| BGE-M38-bit · published params · source | 59.6 | ~3 GB | 8-bit · published params |
Runs well within 640 GB of aggregate GPU memory · planning estimates, not measured speed
Measurements
Published tokens per second, exactly as measured; never estimated.
| DeepSeek V4 Flash FP8 | 39 per user (32 concurrent) | FP8 (sgl-project/DeepSeek-V4-Flash-FP8) | SGLang with EAGLE speculative decoding (docker lmsysorg/sglang:deepseek-v4-hopper) | Lambda inference models · DeepSeek-V4-Flash (vendor benchmark) |
|---|
Specifications
19 sourced fields; the source is named once per group, and a field from another source links to it. Official specifications
Marker: the best LiveBench-scored text model that runs well here. No other text model would run well with the next memory size. Q4 planning estimate, not a measured fit.
Selected configuration
Latest check · Manufacturer source
- Memory capacity
- 640 GB
- Memory type
- 8 × 80 GB HBM3
- Chip / processor
- 8 × NVIDIA H100
- Host CPUs
- 2 × Intel Xeon 8480C
- CPU cores (two CPUs combined)
- 112
- Installed storage (system specification)
- OS: 2 × 1.92 TB NVMe SSD (RAID 1); cache: 8 × 3.84 TB NVMe SSD (RAID 0)
Family specification
Latest check · Manufacturer source
- Storage options (family)
- OS: 2 × 1.92 TB NVMe M.2 SSD in RAID 1
- GPU interconnect
- NVLink GPU-to-GPU: 900 GB/s
- Network connections (system specification)
- 4 × OSFP / 8 × ConnectX-7 400 Gbps · 2 × dual-port ConnectX-7 · BMC 1 GbE RJ45 · management 100 GbE + 10 GbE
- Maximum system input power (specification)
- 10,200 W
- Dimensions (cm)
- 35.6 (H) × 48.23 (W) × 89.71 (D), max
- Weight
- 130.45 kg
- Operating system
- DGX OS (Ubuntu server distribution)
Eight 80 GB GPUs provide 640 GB in aggregate, not one 640 GB address space. Sharding, interconnect and runtime configuration are required for models exceeding one GPU.
Photo: Official manufacturer image. · Manufacturer source
Memory bandwidth is not measured model speed; verify runtime and checkpoint compatibility. Hardware specification reviewed: 2026-10-13.
How estimates work
- Candidates
- Reviewed pages use published parameter counts; automatic text candidates use official repository tensor counts (packed quantized weights and unsupported architectures excluded).
- Planning
- Q4 weights + 20% allowance + reserve: 25% of unified or system memory (at least 2 GB, at most 8 GB; 16 GB here) or 2 GB per dedicated GPU. Adjustable in “Adjust the estimate”. No KV cache is calculated from context.
- Fit tiers
- Runs well: the estimate fits the fast memory (VRAM, or unified memory within the macOS GPU share). Runs slowly: within the machine's memory but outside the fast part; a graphics card offloads layers to PC RAM (the listed RAM, or 32 GB assumed). Doesn't run: above both. Tiers are planning estimates, never a speed.
- Vision
- Vision-language models use the Q4 planning estimate from the parameter count (published by the model page, or read from Hugging Face). Classic vision models (classify, detect, segment) use a 16-bit estimate. No vision benchmark is tracked, so rows carry no score.
- Media and embeddings
- Media rows use documented execution paths (exact hardware or platform) and publisher minimums; a published requirement above this memory moves a model to the next tier. Without either, a 16-bit estimate from the published parameter count (× 1.2 + 2 GB), labelled as such. Embeddings use 8-bit weights + 20% + 2 GB.
Local AI basics: parameters, Q4 and memory · Full model catalog · Data and methodology
