Home / Workstations / P3 Tower / NVIDIA RTX A2000 12 GB
Lenovo P3 Tower
NVIDIA RTX A2000 12 GB · 32 GB system RAM
Hardware
- Family
- Lenovo P3 Tower
- GPU
- NVIDIA RTX A2000 12 GB
- GPU count
- 1
- VRAM
- 12.0 GB advertised
- System RAM
- 32 GB installed
- CPU
- Not stated on the matched listing
What AI can this machine run?
Default below is Q4 at 8K context, labeled. Statuses come from calculator v1.0.0 (single GPU) or Phase 8 (multi-GPU). System RAM is checked against the Phase 9 resident-RAM rule.
Fully fitting at Q4/8K: 43 published models
Fully fitting
- Qwen2.5 0.5B Instruct Fits
- Qwen3 0.6B Fits
- Gemma 3 1B Instruct Fits
- Llama 3.2 1B Instruct Fits
- Qwen2.5 1.5B Instruct Fits
- Qwen2.5-Coder 1.5B Instruct Fits
- Falcon 3 1B Instruct Fits
- SmolLM2 1.7B Instruct Fits
- DeepSeek R1 Distill Qwen 1.5B Fits
- Qwen3 1.7B Fits
- Gemma 2 2B Instruct Fits
- SmolLM3 3B Fits
- Qwen2.5 3B Instruct Fits
- Llama 3.2 3B Instruct Fits
- Falcon 3 3B Instruct Fits
- Phi-3.5 Mini Instruct Fits
- Phi-3 Mini 128K Instruct Fits
- Phi-4 Mini Reasoning Fits
- Phi-4 Mini Instruct Fits
- Qwen3 4B Fits
- Gemma 3 4B Instruct Fits
- Gemma 4 E2B Instruct Fits
- Code Llama 7B Instruct Fits
- StarCoder2 7B Fits
Limited context
- GLM-4 9B Chat Limited context
- Gemma 4 12B Instruct Limited context
- Gemma 3 12B Instruct Limited context
- Code Llama 13B Instruct Limited context
- Phi-3 Medium 128K Instruct Limited context
- Phi-4 Reasoning Limited context
- DeepSeek R1 Distill Qwen 14B Limited context
- Phi-4 Limited context
- Qwen2.5 14B Instruct Limited context
- Qwen3 14B Limited context
Offload required
- StarCoder2 15B Offload
- GPT-OSS 20B Offload
- Magistral Small Offload
- Mistral Small 24B Instruct Offload
- Mistral Small 3.2 24B Instruct Offload
- Gemma 4 26B-A4B Instruct Offload
- Gemma 2 27B Instruct Offload
- Gemma 3 27B Instruct Offload
- Qwen3-Coder 30B-A3B Instruct Offload
- Qwen3 30B-A3B Offload
- Gemma 4 31B Instruct Offload
- DeepSeek R1 Distill Qwen 32B Offload
- Qwen2.5 32B Instruct Offload
- Qwen2.5-Coder 32B Instruct Offload
- QwQ 32B Offload
- Qwen3 32B Offload
- DeepSeek Coder 33B Instruct Offload
- Code Llama 34B Instruct Offload
Does not fit
- Mixtral 8x7B Instruct No
- DeepSeek R1 Distill Llama 70B No
- Hermes 3 Llama 3.1 70B No
- Llama 3.1 70B Instruct No
- Llama 3.3 70B Instruct No
- Qwen2.5 72B Instruct No
- GLM-4.5 Air No
- GPT-OSS 120B No
- Mixtral 8x22B Instruct No
- Qwen3 235B-A22B No · RAM short
- GLM-4.5 No · RAM short
Can it run a selected model?
- Fit
- Fits
- Required VRAM
- 3.0 GB
- Usable GPU memory
- 10.8 GB
- RAM
- 32 GB installed / 16.0 GB required
Performance evidence
No published Phase 5 prediction for this GPU at Q4, or no model selected.
Buy this configuration
New condition only. Storage/OS variants share this AI configuration. Merchant: Amazon.
No fresh eligible Amazon offer is currently attached.
Equivalent custom build
Same selected AI workload via the existing /build/ engine. Arithmetic only.
Comparison is shown only when this workstation has a fresh price and Phase 9 returns a fully priced VALIDATED build for the same model/quant/context.