FitMyLLM
NVIDIA· Blackwell 2.0

NVIDIA RTX PRO 6000D Blackwell Max-Q

Running LLMs on the RTX PRO 6000D Blackwell Max-Q — the long read: which models fit at which quantisation, and the settings worth changing. · Or what a budget buys

VRAM
96 GB
Flagship
Bandwidth
1790
GB/s
Models Q4
422/498
85%
7B Q4 speed
~205
Blazing
Model coverage at Q485% of all
Estimated speed· By model size at Q4

Average speeds at Q4 quantization. Actual performance varies by model architecture and context length.

3B
~477
tok/s
7B
~205
tok/s
14B
~102
tok/s
32B
~45
tok/s
70B
~20
tok/s
Measured rig reports

We rent the machine and time every model on it: decode, VRAM peak, concurrency, watts and cost per million tokens. Yours may already be one of them — and two are free to read in full.

See the reports →
Specifications
VRAM
96 GB
Bandwidth
1790 GB/s
FP16 Compute
110.1 TFLOPS
TDP
300W
Memory
GDDR7
Architecture
Blackwell 2.0
CUDA Cores
24,064
Tensor Cores
752
PCIe
Gen 5 x16
386
Fast models · >30 tok/s
Real-time chat speed
422
Usable · >10 tok/s
Comfortable for all tasks
422
Total compatible
Fit in VRAM at Q4
Rent it in the cloud

Buying RTX PRO 6000D Blackwell Max-Q costs $15–$40k and isn’t practical for most teams. Spin one up by the hour instead:

Spin up in ~60s. Pay by the second. Cancel anytime.

Some links are affiliate links — we may earn a small commission at no extra cost to you. This helps keep FitMyLLM free and independent.

Compatible models· 422
S
Falcon-H1R Tiny 90M0.09B
Falcon·256K ctx· chat· reasoning
15911
tok/s · 1% VRAM
›
S
SmolLM2 135M0.135B
Smollm·2K ctx· chat
10607
tok/s · 1% VRAM
›
S
nomic-embed-text-v1.5 100M0.14B
Embedding·8K ctx· chat
10229
tok/s · 1% VRAM
›
S
GPT-2 124M0.14B
Gpt2·1K ctx· chat
10229
tok/s · 1% VRAM
›
S
SmolVLM 256M0.256B
Smollm·8K ctx· chat· vision
5594
tok/s · 1% VRAM
›
S
Gemma 3 270M0.27B
Gemma·32K ctx· chat
5304
tok/s · 1% VRAM
›
S
Snowflake Arctic Embed M v2.00.305B
Embedding·8K ctx· embedding· multilingual
4695
tok/s · 1% VRAM
›
S
bge-large-en-v1.5 335M0.335B
Embedding·1K ctx· chat
4275
tok/s · 1% VRAM
›
S
mxbai-embed-large-v10.335B
Embedding·1K ctx· embedding
4275
tok/s · 1% VRAM
›
S
Snowflake Arctic Embed L0.335B
Embedding·1K ctx· embedding
4275
tok/s · 1% VRAM
›
S
LFM2 350M0.35B
Lfm·125K ctx· chat· tool_use
4091
tok/s · 1% VRAM
›
S
SmolLM2 360M0.36B
Smollm·8K ctx· chat
3978
tok/s · 1% VRAM
›
S
GPT-2 Medium 345M0.38B
Gpt2·1K ctx· chat
3768
tok/s · 1% VRAM
›
S
Qwen 2.5 0.5B0.5B
Qwen·32K ctx· chat
2864
tok/s · 1% VRAM
›
S
SmolVLM 500M0.5B
Smollm·8K ctx· chat· vision
2864
tok/s · 1% VRAM
›
S
Falcon-H1 0.5B0.52B
Falcon·128K ctx· chat
2754
tok/s · 1% VRAM
›
S
BGE-M30.568B
Embedding·8K ctx· embedding
2521
tok/s · 1% VRAM
›
S
Snowflake Arctic Embed L v2.00.568B
Embedding·8K ctx· embedding· multilingual
2521
tok/s · 1% VRAM
›
S
Qwen3-Embedding 0.6B0.6B
Embedding·32K ctx· embedding· multilingual
2387
tok/s · 1% VRAM
›
S
Falcon-H1R Tiny 0.6B0.6B
Falcon·32K ctx· chat· reasoning
2387
tok/s · 1% VRAM
›
S
Falcon Perception 0.6B0.6B
Falcon·4K ctx· vision
2387
tok/s · 1% VRAM
›
S
Qwen 1.5 0.5B0.62B
Qwen·32K ctx· chat
2310
tok/s · 1% VRAM
›
S
LFM2 700M0.74B
Lfm·125K ctx· chat· tool_use
1935
tok/s · 1% VRAM
›
S
Qwen3 0.6B0.75B
Qwen·32K ctx· chat· reasoning
1909
tok/s · 1% VRAM
›
S
GPT-2 Large 774M0.81B
Gpt2·1K ctx· chat
1768
tok/s · 1% VRAM
›
S
Qwen 3.5 0.8B0.87B
Qwen·256K ctx· chat· coding· multilingual
1646
tok/s · 1% VRAM
›
S
InternVL3 1B0.94B
Other·32K ctx· chat· vision
1523
tok/s · 1% VRAM
›
S
MiniCPM5 1B1.08B
Minicpm·128K ctx· chat· reasoning· multilingual
1326
tok/s · 1% VRAM
›
S
K2 Horizon 0.9B1.08B
Other·128K ctx· chat· coding· reasoning
1326
tok/s · 1% VRAM
›
S
TinyLlama 1.1B1.1B
Llama·2K ctx· chat
1302
tok/s · 1% VRAM
›
S
LFM2.5-1.2B-Thinking1.2B
Lfm·122K ctx· chat· reasoning· tool_use
1193
tok/s · 1% VRAM
›
S
Llama-3.2-1B1.2B
Llama·4K ctx· chat
1193
tok/s · 1% VRAM
›
S
LFM2 1.2B1.2B
Lfm·125K ctx· chat· tool_use· multilingual
1193
tok/s · 1% VRAM
›
S
Zamba2 1.2B1.2B
Other·4K ctx· chat
1193
tok/s · 1% VRAM
›
S
Ling 3.0 Tiny7.89BMoE
Other·128K ctx· chat· coding· reasoning
1102
tok/s · 6% VRAM
›
S
EXAONE-4.0-1.2B1.3B
Exaone·64K ctx· chat
1102
tok/s · 1% VRAM
›
S
OPT 1.3B1.3B
Opt·2K ctx· chat
1102
tok/s · 1% VRAM
›
S
MiniCPM-V 4.61.3B
Other·256K ctx· chat· vision
1102
tok/s · 1% VRAM
›
S
DeepSeek Coder 1.3B1.35B
Deepseek·16K ctx· coding
1061
tok/s · 1% VRAM
›
S
Phi-1 1.3B1.42B
Phi·2K ctx· coding
1008
tok/s · 1% VRAM
›
Next step

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your RTX PRO 6000D Blackwell Max-Q.

What this card is worth

RTX PRO 6000D Blackwell Max-Q holds 422 of the models in our catalogue and is, in practice, a IQ4_XS card — the largest it takes is dots.llm1 142B-A14B at IQ4_XS.

Tokens/sec per $100
—
8B at Q4_K_M, so cards compare like for like
VRAM per $100
—
what memory costs on this card
The biggest it takes
dots.llm1 142B-A14B142.8B · IQ4_XS86.0 GB90 tok/sest
WizardLM 2 8x22B141B · Q4_K_S85.7 GB31 tok/sest
Mixtral-8x22B140.6B · Q4_K_S85.5 GB31 tok/sest
DBRX 132B132B · Q4_K_M83.8 GB32 tok/sest
Mistral Medium 3.5128B · Q4_K_M82.1 GB9 tok/sest
Ling 3.0 Flash127.49B · Q4_K_M80.5 GB227 tok/sest
Nemotron 3.5 Super VL 120B-A12B124.45B · Q4_K_M78.6 GB96 tok/sest
Pixtral Large 124B124B · Q4_K_M79.7 GB9 tok/sest
Against cards of similar memory
Radeon Instinct MI300128 GB$12,0005.8 tok/s per $100
Radeon Instinct MI250128 GB$12,0002.9 tok/s per $100
Instinct MI250X128 GB$10,0003.5 tok/s per $100
M1 Ultra (128GB)96 GB$4,9991.4 tok/s per $100
RTX PRO 6000D Blackwell Max-Q96 GB——

Neighbours in memory rather than in price: memory decides whether a card can do the job at all, so two cards of the same size at different prices is the comparison you are making. Ordered by memory, then bandwidth — not by the value column, which is worked out from bandwidth and price alone and therefore rewards a cheap card whatever its software stack does to that bandwidth in practice. Read it as one input, not as a ranking.

Device under test

NVIDIA RTX PRO 6000D Blackwell Max-Q — 96 GB VRAM.

RTX PRO 6000D Blackwell Max-Q spec
Brand
NVIDIA
VRAM
96 GB GDDR7
Bandwidth
1790 GB/s
FP16 Compute
110.1 TFLOPS
FP32 Compute
110.1 TFLOPS
CUDA Cores
24,064
Tensor Cores
752
TDP
300 W
Architecture
Blackwell 2.0
AI capability
422/ 498 models @ Q4

With 96 GB VRAM and 1790 GB/s bandwidth, this GPU handles models up to 132B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~205 tok/s.

Rent it — live prices
Checking what the clouds are charging…
Top models for RTX PRO 6000D Blackwell Max-Q
422 fit · showing 20
ModelSizeVRAM Q4tok/sAvg
DBRX 132B132B81.2 GB4046.3
Mistral Medium 3.5128B78.7 GB1144.5
Ling 3.0 Flash127.49B78.4 GB28157.0
Nemotron 3.5 Super VL 120B-A12B124.45B76.6 GB11956.6
Pixtral Large 124B124B76.3 GB1284.4
Nemotron 3 Super 120B-A12B123.61B76.0 GB11953.5
Mistral-Large 123B123B75.7 GB1233.5
Devstral 2 123B123B75.7 GB1235.9
Qwen 3.5 122B A10B122B75.1 GB14353.8
Nemotron 3 Super 120B120B73.8 GB11952.3
Mistral Small 4 119B119B73.2 GB22050.2
GPT-OSS 120B117B72.0 GB281153.1
Command A 111B111B68.3 GB1380.8
GLM 4.5 Air110B67.7 GB11951.0
Qwen 1.5 110B110B67.7 GB1333.4
Llama 4 Scout 17B-16E109B67.1 GB8488.9
Cogito v2 109B MoE109B67.1 GB84—
Ling 2.6 Flash107.49B66.2 GB19442.6
Sarvam 105B105B64.7 GB1448.0
Command-R+ 104B104B64.1 GB1452.7