VOOZH about

URL: https://willitrunai.com/gpus/b200-180gb

โ‡ฑ AI Models for NVIDIA B200 180GB โ€” What Runs on 180GB VRAM


NVIDIA

NVIDIA B200 180GB

Blackwell DatacenterDatacenterBlackwellNVLINKCUDA
180GB
VRAM
8kGB/s
Bandwidth
2.3kTFLOPS
FP16 Compute
4.5kTOPS
INT8 Inference
$30,000 MSRP
NVIDIA B200 180GBCategory AvgAMD Instinct MI325X 256GB

Operating mode

Choose the operating mode for this hardware

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.

See Full AI Tier List for NVIDIA B200 180GB โ†’

About this GPU for AI

The NVIDIA B200 is the flagship Blackwell datacenter GPU, delivering 180 GB of HBM3e and 2,250 TFLOPS of FP16 compute โ€” roughly 2.3x the compute of an H100 at over twice the VRAM. Its new fourth-generation Tensor Cores add FP4 support, enabling up to 4,500 TOPS for FP4 inference, and the Blackwell architecture introduces a second-generation Transformer Engine. A single B200 can serve 70B models at FP16 with headroom for large batch sizes and long context windows, making it suitable for high-throughput production inference. At approximately 1,000W TDP, it targets next-generation liquid-cooled infrastructure.

Beyond LLMs

AI Capability Matrix

What AI tasks this GPU can handle โ€” from text generation to image and video creation.

CapabilityStatusRepresentative ModelDetail
LLM Chat (7B)Runs nativelyLlama 3.1 8B Q4โ€”
LLM Coding (30B)Runs nativelyQwen 3 30B Q4โ€”
LLM Large (70B)Runs nativelyLlama 3.1 70B Q4โ€”
Image Gen (SDXL)Runs nativelySDXL 1.0 FP16~100ms per image
Image Gen (Flux)Runs nativelyFlux.1 Dev FP16~600ms per image
Image Gen (SD 3.5)Runs nativelySD 3.5 Large FP16~700ms per image
Video Short (25f)Runs nativelyLTX Video 2B~100ms/frame
Video Long (100f)Runs nativelyWan Video 14B~300ms/frame
hbm-memorymassive-vramhigh-bandwidthblackwell-architecturebest-in-classpower-hungry

Specifications

Compute
FP162250 TFLOPS
INT84500 TOPS
ArchitectureBlackwell
Memory
VRAM180 GB
Bandwidth8000 GB/s
General
FamilyBlackwell Datacenter
SegmentDatacenter
InterconnectNVLINK
Compute PlatformCUDA
MSRP$30,000

Key Features

180 GB HBM3e โ€” largest memory capacity in the B200 lineup8,000 GB/s memory bandwidth2,250 TFLOPS FP16 / 4,500 INT8 TOPS / FP4 Tensor Core support2nd-gen Transformer Engine for FP8 and FP4 inferenceNVLink 5.0 with 1.8 TB/s per-GPU bandwidth for multi-GPU scaling~1,000W TDP โ€” requires liquid or next-gen air cooling

For AI Workloads

Strengths
  • 180 GB HBM3e handles 70B models at FP16 and 405B+ models with Q4 on a single card
  • 8 TB/s bandwidth is among the highest available, enabling fast token generation at large batch sizes
  • FP4 Tensor Cores deliver up to 2.3x higher inference throughput vs. H100 FP8
  • NVLink 5.0 enables efficient 8-GPU HGX B200 clusters with 1.44 TB pooled memory
Considerations
  • ~1,000W TDP demands liquid cooling infrastructure โ€” not compatible with legacy H100 SXM racks
  • Extremely high cost โ€” list pricing well above H100, with significant waitlists
  • Software ecosystem still maturing โ€” TensorRT-LLM and vLLM FP4 support launched recently
  • Overkill for serving models below 30B parameters; ROI requires high-utilization production workloads

Architecture

Blackwell

Blackwell is NVIDIA's fifth-generation RTX architecture, built on TSMC's 4NP process. It introduces 5th-generation Tensor Cores with native FP4 precision support, enabling double the inference throughput per watt compared to Ada Lovelace's FP8 operations. Key innovations include the Neural Rendering Pipeline for AI-driven shading and the debut of GDDR7 memory in consumer GPUs.

AI Relevance

FP4 Tensor Cores deliver the highest tokens-per-watt efficiency in any consumer architecture. Native FP4 quantization means models can run at lower precision with minimal quality loss, effectively doubling the effective VRAM for model weights.

Process: TSMC 4NPPlatform: CUDATensor Cores: Gen 5Precisions: FP32, FP16, BF16, FP8, FP4, INT8, INT4

Buying advice

Should you buy NVIDIA B200 180GB for local AI?

Excellent choice for local AI

Runs 39 of 50 top models well โ€” a strong all-rounder for local inference.

180.0 GB

VRAM

$30,000

MSRP

$167/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 4 additional models that do not fit on the current setup.

Want more headroom? AMD Instinct MI325X 256GB (256.0 GB VRAM) is the next step up.

Recommendations by Workload

Chat

S

Mistral Small 4 119B

This 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.

Decode 292.9 tok/s ยท 256K ctx ยท llama.cppEST.
94.2 GB / 180.0 GB VRAM

Coding

S

Devstral 2 123B Instruct

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, lm-studio.

Decode 97.4 tok/s ยท 256K ctx ยท llama.cppEST.
99.3 GB / 180.0 GB VRAM

Agentic Coding

S

Devstral 2 123B Instruct

This 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.

Decode 97.4 tok/s ยท 256K ctx ยท llama.cppEST.
104.7 GB / 180.0 GB VRAM

Reasoning

S

Devstral 2 123B Instruct

This 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.

Decode 97.4 tok/s ยท 256K ctx ยท llama.cppEST.
99.3 GB / 180.0 GB VRAM

RAG

S

Qwen 3.5 122B A10B

This 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.

Decode 270.2 tok/s ยท 131K ctx ยท llama.cppEST.
98.2 GB / 180.0 GB VRAM

Full Model Compatibility

๐Ÿ‘ Mistral
Devstral 2 123B Instruct
S97
123B99.3 GB97 tok/s256K ctx
dense
๐Ÿ‘ DeepSeek
DeepSeek V4 Flash
S96
284B178.2 GB145 tok/s38K ctx
moe
๐Ÿ‘ Alibaba
Qwen 3.5 122B A10B
S96
122B95.8 GB270 tok/s131K ctx
moe
๐Ÿ‘ Mistral
Mistral Small 4 119B
S94
119B96.9 GB293 tok/s256K ctx
moe
๐Ÿ‘ OpenAI
GPT-OSS 120B
S94
117B95.2 GB102 tok/s131K ctx
dense
๐Ÿ‘ Mistral AI
Pixtral Large 124B
S93
124B99.9 GB97 tok/s131K ctx
dense
๐Ÿ‘ Cohere
Command A 111B
S93
111B90.5 GB108 tok/s262K ctx
dense
๐Ÿ‘ Alibaba
Qwen 3 235B A22B
S92
235B165.1 GB137 tok/s99K ctx
moe
๐Ÿ‘ Alibaba
Qwen 2.5 VL 72B
S90
72B67.7 GB166 tok/s33K ctx
dense
MiniMax M2.7
S90
230B163.0 GB156 tok/s88K ctx
moe
๐Ÿ‘ Alibaba
Qwen3-Coder-Next
S90
80B69.2 GB454 tok/s256K ctx
moe
๐Ÿ‘ Mistral
Leanstral 119B A6B
S90
119B100.3 GB269 tok/s161K ctx
moe
๐Ÿ‘ Alibaba
Qwen3-Coder 30B A3B Instruct
S90
30.5B39.0 GB1016 tok/s256K ctx
moe
๐Ÿ‘ Alibaba
Qwen 3.6 35B A3B
S89
35B44.4 GB854 tok/s262K ctx
+1moe
๐Ÿ‘ Alibaba
Qwen 3.5 27B
S89
27B38.5 GB378 tok/s131K ctx
dense
๐Ÿ‘ Alibaba
Qwen3-VL 30B A3B Instruct
S89
30B38.7 GB1051 tok/s256K ctx
moe
๐Ÿ‘ Alibaba
Qwen 3.6 27B
S89
27B36.3 GB275 tok/s262K ctx
+1dense
๐Ÿ‘ Alibaba
Qwen 3.5 35B A3B
S88
35B41.7 GB929 tok/s131K ctx
moe
S88
32B42.3 GB374 tok/s131K ctx
dense
๐Ÿ‘ Mistral
Magistral Small 2507
S88
24B36.0 GB336 tok/s131K ctx
dense
๐Ÿ‘ Mistral
Devstral Small 2 24B Instruct
S87
24B36.0 GB336 tok/s256K ctx
dense
๐Ÿ‘ Alibaba
Qwen 3 30B A3B
S87
30.5B39.0 GB1016 tok/s131K ctx
moe
๐Ÿ‘ NVIDIA
Nemotron 3 Nano 30B
S87
30B39.6 GB395 tok/s131K ctx
dense
S87
9B26.6 GB126 tok/s131K ctx
dense
S86
14B29.9 GB196 tok/s131K ctx
dense
๐Ÿ‘ Mistral
Devstral Small 1.1
S86
24B36.0 GB336 tok/s131K ctx
dense
๐Ÿ‘ Microsoft
Phi-4-reasoning-plus 14B
S85
14.7B30.9 GB206 tok/s33K ctx
dense
๐Ÿ‘ Google
Gemma 4 31B
A85
30.7B52.3 GB234 tok/s156K ctx
dense
A85
8B26.0 GB112 tok/s131K ctx
dense
๐Ÿ‘ OpenAI
GPT-OSS 20B
A84
21B34.2 GB1290 tok/s128K ctx
moe
๐Ÿ‘ NVIDIA
Nemotron Cascade 2 30B A3B
A84
30B40.1 GB1039 tok/s262K ctx
moe
A83
4B23.5 GB56 tok/s131K ctx
dense
๐Ÿ‘ LG AI
EXAONE 4.0 32B
A82
32B42.3 GB372 tok/s131K ctx
dense
๐Ÿ‘ Google
Gemma 4 26B A4B
A81
25.2B37.9 GB1091 tok/s256K ctx
moe
๐Ÿ‘ Mistral
Ministral 3 14B
A80
14B29.9 GB196 tok/s262K ctx
multimodal
๐Ÿ‘ NVIDIA
Nemotron Nano 8B
A80
8B25.7 GB112 tok/s131K ctx
dense
๐Ÿ‘ Microsoft
Phi-4 Mini Reasoning 4B
A80
3.8B22.7 GB53 tok/s131K ctx
dense
๐Ÿ‘ Jina AI
Jina Embeddings v3
A74
0.57B22.0 GB8 tok/s8K ctx
dense
0.57B21.2 GB8 tok/s8K ctx
dense
๐Ÿ‘ Alibaba
Qwen 3.5 397B A17B
F0
397B263.9 GB41 tok/s4K ctx
moe
1000B636.3 GB6 tok/s4K ctx
moe
1000B636.3 GB6 tok/s4K ctx
+1moe
๐Ÿ‘ DeepSeek
DeepSeek V4 Pro
F0
1600B882.8 GB4 tok/s4K ctx
moe
754B497.9 GB7 tok/s4K ctx
moe
744B491.8 GB7 tok/s4K ctx
moe
๐Ÿ‘ DeepSeek
DeepSeek V3.2
F0
671B428.7 GB11 tok/s4K ctx
moe
๐Ÿ‘ Alibaba
Qwen3-Coder 480B A35B Instruct
F0
480B314.6 GB24 tok/s4K ctx
moe
๐Ÿ‘ DeepSeek
DeepSeek Coder V2 236B
F0
236B221.5 GB76 tok/s5K ctx
moe
๐Ÿ‘ DeepSeek
DeepSeek R1 671B
F0
671B487.8 GB8 tok/s4K ctx
moe
๐Ÿ‘ DeepSeek
DeepSeek V3.1 671B
F0
671B487.8 GB8 tok/s4K ctx
moe

Just out of reach

Models you could run with an upgrade

High-quality models that need a bit more memory

Image & Video Generation

Diffusion Model Compatibility

52 of 52 models can generate images or video on your NVIDIA B200 180GB

ModelMax ResolutionGen TimeGrade
SD TurboImage512ร—5120msS
Stable Diffusion 1.5Image512ร—7680msS
Realistic Vision v5.1Image512ร—7680msS
DreamShaper 8Image512ร—7680msS
LCM DreamShaper v7Image512ร—7680msS
PixArt-SigmaImage1024ร—1024100msS
FramePack I2VVideo1280ร—720200ms/frameS
SDXL TurboImage512ร—5120msS
SDXL LightningImage1024ร—10240msS
Stable Diffusion XL 1.0Image1024ร—1024100msS
Playground v2.5Image1024ร—1024200msS
RealVisXL v5.0Image1024ร—1024100msS
DreamShaper XLImage1024ร—1024100msS
Juggernaut XL v9Image1024ร—1024100msS
Animagine XL 3.1Image1024ร—1024100msS
Pony Diffusion V6 XLImage1024ร—1024100msS
Animagine XL 4.0Image1024ร—1024100msS
Illustrious XLImage1024ร—1024100msS
Wan Video 2.1 1.3BVideo480ร—832100ms/frameS
Stable Diffusion 3.5 MediumImage1024ร—1024200msS
Flux.2 Klein 4BImage1024ร—10240msS
LTX Video 2BVideo1280ร—720100ms/frameS
KolorsImage1024ร—1024300msS
Stable CascadeImage1024ร—1024300msS
AuraFlow v0.3Image1536ร—1536600msS
Stable Diffusion 3.5 LargeImage1024ร—1024700msS
Stable Diffusion 3.5 Large TurboImage1024ร—1024100msS
CogVideoX 2BVideo720ร—480100ms/frameS
HunyuanVideoVideo720ร—1280200ms/frameS
ChromaImage1024ร—1024100msS
Z-Image TurboImage1536ร—1536100msS
Flux.1 DevImage1024ร—1024600msS
Flux.1 SchnellImage1024ร—1024100msS
LTX Video 13BVideo1280ร—720200ms/frameS
Flux.1 Kontext DevImage1024ร—1024700msS
AnimateDiff v1.5.3Video512ร—768100ms/frameS
Cosmos Diffusion 7BVideo1024ร—576200ms/frameS
CogVideoX 5BVideo720ร—480200ms/frameS
Wan2.2 TI2V 5BVideo832ร—480200ms/frameS
Flux.2 Klein 9BImage1024ร—1024100msS
Flux.1 Fill DevImage1024ร—1024600msS
Mochi 1 PreviewVideo848ร—480200ms/frameS
HunyuanVideo 1.5Video720ร—1280200ms/frameS
Helios 14BVideo1280ร—720200ms/frameS
SkyReels V2 14BVideo1280ร—720200ms/frameS
Wan Video 2.1 14BVideo720ร—1280200ms/frameS
Wan Video 2.2 14BVideo720ร—1280200ms/frameS
Qwen ImageImage1024ร—1024200msS
Qwen Image EditImage1024ร—1024200msS
Flux.2 DevImage1024ร—1024~6.2sS
MAGI-1Video1280ร—720300ms/frameS
HunyuanImage 3.0Image1024ร—1024400msB

Image models estimated at 1024ร—1024 (28 steps, FP16). Video models estimated at 768ร—512 (25 frames, 30 steps, FP16). Actual performance varies with runtime and system load.

Multi-GPU scaling

NVIDIA B200 180GB โ€” Up to 8ร— via NVLink

Scale out with multiple GPUs for larger models. NVLink provides 1800 GB/s inter-GPU bandwidth with 7% overhead.

ConfigEffective memoryModels that fitEst. bandwidth
1ร— NVIDIA180 GB359/3748,000 GB/s
2ร— NVIDIA360 GB364/37414,880 GB/s
4ร— NVIDIA720 GB373/37429,760 GB/s
8ร— NVIDIA1440 GB374/37459,520 GB/s

Model counts use default quantization at coding workload settings. Multi-GPU scaling factor: 0.93ร— per additional GPU.

Upgrade paths

Upgrade from NVIDIA B200 180GB

See what you unlock with more powerful hardware

Upgrade options

Upgrade options

๐Ÿ‘ NVIDIA
8ร— NVIDIA B200 180GBMulti-GPU
8 ร— 180 GB = 1440 GB effectivevia NVLink (1800 GB/s)
B
Unlocks 15 additional models that do not fit on the current setup.Unlocks Qwen 3.5 397B A17B, Kimi K2.5, Kimi K2.6+12 more ยท +171% faster avg

Unlocks 15 additional models that do not fit on the current setup.

Lifts average decode speed across fitting models by about 171%.

NVLink gives this scale-out path a cleaner inter-GPU story than PCIe-only builds.

~$30,000 MSRP

AMD Instinct MI325X 256GBNext step up
256 GB VRAM (+76)
B
Unlocks 4 additional models that do not fit on the current setup.Unlocks Qwen 3.5 397B A17B, Llama 4 Maverick 17B 128E, Llama 3.1 405B+1 more

Unlocks 4 additional models that do not fit on the current setup.

~$20,000 MSRP

AMD Instinct MI350X 288GBBest value
288 GB VRAM (+108)
B
Unlocks 5 additional models that do not fit on the current setup.Unlocks Qwen 3.5 397B A17B, Qwen3-Coder 480B A35B Instruct, Llama 4 Maverick 17B 128E+2 more

Unlocks 5 additional models that do not fit on the current setup.

~$8,000 MSRP

Frequently Asked Questions

Compare with similar

NVIDIA B200 180GB vs H100 NVL 188GBNVIDIA B200 180GB vs AMD Instinct MI300X 192GBNVIDIA B200 180GB vs NVIDIA GB200 192GB
Compare this GPUCompare with another GPU โ†’