FitMyLLM
▸ NVIDIA· FERMI

NVIDIA Quadro 5000 SDI

VRAM
3 GB
BUDGET
BANDWIDTH
120
GB/S
MODELS Q4
52/331
16%
7B Q4 SPEED
~14
USABLE
▸ MODEL COVERAGE @ Q416% OF ALL
▸ ESTIMATED SPEED· BY MODEL SIZE @ Q4

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

3B
~32
TOK/S
7B
—
3.9GB NEEDED
14B
—
7.9GB NEEDED
32B
—
18.0GB NEEDED
70B
—
39.4GB NEEDED
▸ SPECIFICATIONS
VRAM
3 GB
BANDWIDTH
120 GB/s
FP16 COMPUTE
0.7 TFLOPS
TDP
172W
MEMORY
GDDR5
ARCHITECTURE
Fermi
CUDA CORES
352
PCIE
Gen 2 x16
52
FAST MODELS · >30 TOK/S
Real-time chat speed
52
USABLE · >10 TOK/S
Comfortable for all tasks
52
TOTAL COMPATIBLE
Fit in VRAM at Q4
▸ DON’T WANT TO BUY?

Test Quadro 5000 SDI (or anything bigger) without committing. Pay by the second, cancel anytime.

Rent on RunPod →Or Vast.ai →

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· 52
S
nomic-embed-text-v1.5 100M0.1B
EMBEDDING·8K CTX· CHAT
960
TOK/S · 22% VRAM
›
S
GPT-2 124M0.124B
GPT2·1K CTX· CHAT
774
TOK/S · 23% VRAM
›
S
SmolLM2 135M0.135B
SMOLLM·2K CTX· CHAT
711
TOK/S · 23% VRAM
›
S
bge-large-en-v1.5 335M0.335B
EMBEDDING·1K CTX· CHAT
287
TOK/S · 28% VRAM
›
S
mxbai-embed-large-v10.335B
EMBEDDING·1K CTX· EMBEDDING
287
TOK/S · 28% VRAM
›
S
Snowflake Arctic Embed L0.335B
EMBEDDING·1K CTX· EMBEDDING
287
TOK/S · 28% VRAM
›
S
GPT-2 Medium 345M0.345B
GPT2·1K CTX· CHAT
278
TOK/S · 28% VRAM
›
S
SmolLM2 360M0.36B
SMOLLM·8K CTX· CHAT
267
TOK/S · 28% VRAM
›
S
Falcon-H1 0.5B0.5B
FALCON·128K CTX· CHAT
192
TOK/S · 32% VRAM
›
S
Qwen 1.5 0.5B0.5B
QWEN·32K CTX· CHAT
192
TOK/S · 32% VRAM
›
S
Qwen 2.5 0.5B0.5B
QWEN·32K CTX· CHAT
192
TOK/S · 32% VRAM
›
S
BGE-M30.568B
EMBEDDING·8K CTX· EMBEDDING
169
TOK/S · 33% VRAM
›
S
Qwen3 0.6B0.6B
QWEN·32K CTX· CHAT· REASONING
160
TOK/S · 34% VRAM
›
S
GPT-2 Large 774M0.774B
GPT2·1K CTX· CHAT
124
TOK/S · 38% VRAM
›
S
Qwen 3.5 0.8B0.8B
QWEN·256K CTX· CHAT· CODING· MULTILINGUAL
120
TOK/S · 39% VRAM
›
S
Qwen3.5-0.8B0.9B
QWEN·256K CTX· CHAT
107
TOK/S · 42% VRAM
›
S
Falcon3-1B1B
FALCON·32K CTX· CHAT
96
TOK/S · 44% VRAM
›
S
InternLM2 1B1B
INTERNLM·32K CTX· CHAT
96
TOK/S · 44% VRAM
›
S
TinyLlama 1.1B1.1B
LLAMA·2K CTX· CHAT
87
TOK/S · 46% VRAM
›
S
LFM2.5-1.2B-Thinking1.2B
LFM·122K CTX· CHAT· REASONING· TOOL_USE
80
TOK/S · 49% VRAM
›
S
Llama-3.2-1B1.2B
LLAMA·4K CTX· CHAT
80
TOK/S · 49% VRAM
›
S
DeepSeek Coder 1.3B1.3B
DEEPSEEK·16K CTX· CODING
74
TOK/S · 51% VRAM
›
S
EXAONE-4.0-1.2B1.3B
EXAONE·64K CTX· CHAT
74
TOK/S · 51% VRAM
›
S
OPT 1.3B1.3B
OPT·2K CTX· CHAT
74
TOK/S · 51% VRAM
›
S
Phi-1 1.3B1.3B
PHI·2K CTX· CODING
74
TOK/S · 51% VRAM
›
S
Phi-1.5 1.3B1.3B
PHI·2K CTX· CHAT· CODING
74
TOK/S · 51% VRAM
›
S
Falcon-H1 1.5B1.5B
FALCON·128K CTX· CHAT· CODING
64
TOK/S · 56% VRAM
›
S
GPT-2 XL 1.5B1.5B
GPT2·1K CTX· CHAT
64
TOK/S · 56% VRAM
›
S
Qwen2.5-Coder-1.5B1.5B
QWEN·32K CTX· CHAT· TOOL_USE· CODING
64
TOK/S · 56% VRAM
›
S
Qwen2 Math 1.5B1.5B
QWEN·4K CTX· REASONING
64
TOK/S · 56% VRAM
›
S
Qwen 2.5 1.5B1.5B
QWEN·32K CTX· CHAT· CODING
64
TOK/S · 56% VRAM
›
S
Yi Coder 1.5B1.5B
YI·125K CTX· CODING
64
TOK/S · 56% VRAM
›
S
stablelm-2-1_6b1.6B
STABLELM·4K CTX· CHAT
60
TOK/S · 59% VRAM
›
A
SmolLM2 1.7B1.71B
SMOLLM·8K CTX· CHAT
56
TOK/S · 61% VRAM
›
A
Qwen3 1.7B1.7B
QWEN·32K CTX· CHAT· REASONING
56
TOK/S · 61% VRAM
›
A
Qwen 1.5 1.8B1.8B
QWEN·32K CTX· CHAT
53
TOK/S · 64% VRAM
›
A
Moondream2 1.9B1.9B
OTHER·2K CTX· VISION· CHAT
51
TOK/S · 66% VRAM
›
A
Gemma 1 2B2B
GEMMA·8K CTX· CHAT
48
TOK/S · 68% VRAM
›
A
Granite 3.0 2B2B
GRANITE·128K CTX· CHAT· CODING
48
TOK/S · 68% VRAM
›
A
Granite 3.1 2B2B
GRANITE·128K CTX· CHAT· CODING
48
TOK/S · 68% VRAM
›
▸ NEXT STEP

Get personalized recommendations.

See ranked models with benchmark scores, run commands, and precise speed estimates for your Quadro 5000 SDI.

▸Find models for Quadro 5000 SDICompare with other GPUs
▸ DEVICE UNDER TEST

NVIDIA Quadro 5000 SDI — 3 GB VRAM.

▸ QUADRO 5000 SDI SPEC
BRAND
NVIDIA
VRAM
3 GB GDDR5
BANDWIDTH
120 GB/s
FP16 COMPUTE
0.7 TFLOPS
FP32 COMPUTE
0.7 TFLOPS
CUDA CORES
352
TDP
172 W
ARCHITECTURE
Fermi
▸ AI CAPABILITY
52/ 331 models @ Q4

With 3 GB VRAM and 120 GB/s bandwidth, this GPU handles models up to 2.8B parameters.

Speed ≈ bandwidth / model_size × efficiency. A 7B model at Q4 runs at ~14 tok/s.

§ 01TOP MODELS FOR QUADRO 5000 SDI
52 FIT · SHOWING 20
MODELSIZEVRAM Q4TOK/SAVG
Dolly v2 3B2.8B2.2 GB345.6
StableLM Zephyr 3B2.79B2.2 GB3414.9
Zephyr 3B2.79B2.2 GB3414.4
OPT 2.7B2.7B2.1 GB3628.0
Phi-2 2.7B2.7B2.1 GB3624.1
gemma-2-2b2.6B2.1 GB3722.9
CodeGemma 2B2.51B2.0 GB3822.9
EXAONE Deep 2.4B2.4B2.0 GB4027.1
Qwen3.5-2B2.3B1.9 GB4219.3
Qwen2-VL 2B2.21B1.8 GB4328.3
Gemma 1 2B2B1.7 GB4820.2
Granite 3.0 2B2B1.7 GB4835.8
Granite 3.1 2B2B1.7 GB4837.8
Granite 3.3 2B2B1.7 GB4820.5
Qwen 3.5 2B2B1.7 GB4829.7
Moondream2 1.9B1.9B1.6 GB5126.2
Qwen 1.5 1.8B1.8B1.6 GB5319.6
SmolLM2 1.7B1.71B1.5 GB5616.2
Qwen3 1.7B1.7B1.5 GB5630.1
stablelm-2-1_6b1.6B1.5 GB609.5