← All GPUs
NVIDIA · Datacenter (used) · Pascal (GP102)
NVIDIA Tesla P40
CUDA 13 dropped Pascal, so it is stuck on CUDA 12.x builds; no fan, needs a shroud or server airflow.
VRAM
24 GB
GDDR5
Memory bandwidth
346 GB/s
FP16 compute
not published
Price
Used only
used price not tracked yet
What fits at 4-bit
The 4-bit file plus 8,192 tokens of context and 1.5 GB of overhead, against the VRAM of one card. Speeds are measured medians where someone has published a run at that card count; a blank speed means no run, not a slow one. Machine pages carry estimates for the gaps.
| Model | Needs | 1 card (24 GB) |
|---|---|---|
| Qwen3.5 4B4B · Q4_K_M | 4.3 GB | fits |
| Llama 3.1 8B Instruct8B · Q4_K_M | 7.1 GB | fits |
| Qwen3 8B8.2B · Q4_K_M | 7.3 GB | fits |
| Gemma 4 E4B IT8B · Q4_K_M | 6.5 GB | fits |
| Granite 4.2 8B8.8B · Q4_K_M | 7.9 GB | fits |
| Qwen3.5 9B9B · Q4_K_M | 7.0 GB | fits |
| Gemma 4 12B IT11.95B · Q4_K_M | 8.2 GB | fits |
| Ministral 3 14B Reasoning (2512)13.9B · Q4_K_M | 10.4 GB | fits |
| Phi-414.7B · Q4_K_M | 11.3 GB | fits |
| gpt-oss-20b21B, 3.6B active · MXFP4 (native) | 12.5 GB | fits |
| Devstral Small 2 24B (2512)24B · Q4_K_M | 16.1 GB | fits |
| Qwen3.6 27B27B · Q4_K_M | 17.7 GB | fits |
| Gemma 4 26B-A4B IT25.2B, 3.8B active · Q4_K_M | 17.4 GB | fits |
| Qwen3.8 27B27B · Q4_K_M | 18.2 GB | fits |
| Granite 4.2 30B29.3B · Q4_K_M | 20.3 GB | fits |
| Gemma 4 31B IT30.7B · Q4_K_M | 18.9 GB | fits |
| Qwen3 Coder 30B-A3B Instruct30.5B, 3.3B active · Q4_K_M | 19.5 GB | fits |
| Ornith 1.5 35B-A3B35B, 3B active · Q4_K_M | 21.9 GB | fits |
| Qwen3.6 35B-A3B35B, 3B active · Q4_K_M | 22.4 GB | fits |
Measured speeds
One card
| Model | Quant | Backend | Generation | Prompt | Context | Source |
|---|---|---|---|---|---|---|
| Qwen3.6 35B-A3B | Q3_K_XL | llama.cpp (build not stated, CUDA 12.9 sm_61, FORCE_MMQ) | 38 tok/s | – | 0 | llama.cpp benchmark discussion2026-07 |
| Qwen3.6 35B-A3B | Q3_K_XL | llama.cpp (build not stated, CUDA 12.9 sm_61, FORCE_MMQ) | – | 790 tok/s6,000-token prompt | 0 | llama.cpp benchmark discussion2026-07 |
| Qwen3.6 35B-A3B | Q3_K_XL | llama.cpp (build not stated, CUDA 12.9 sm_61, FORCE_MMQ) | – | 422 tok/s25,000-token prompt | 0 | llama.cpp benchmark discussion2026-07 |
| Qwen3.6 35B-A3B | Q3_K_XL | llama.cpp (build not stated, CUDA 12.9 sm_61, FORCE_MMQ) | – | 216 tok/s78,000-token prompt | 0 | llama.cpp benchmark discussion2026-07 |
Speculative decoding, several users, or long context
Kept apart because each measures a different workload from one person chatting near the start of a conversation.
| Model | Quant | Backend | Generation | Prompt | Context | Measured on | Source |
|---|---|---|---|---|---|---|---|
| Qwen3.6 35B-A3B | Q3_K_XL | llama.cpp (build not stated, CUDA 12.9 sm_61, FORCE_MMQ) | 52 tok/s | 982 tok/s512-token prompt | 0 | Tesla P40 | llama.cpp benchmark discussion2026-07 |
Specifications
- Memory
- 24 GB GDDR5
- FP16 basis
- No tensor cores; NVIDIA lists 12 TFLOPS FP32 and 47 TOPS INT8 only.
- Board power
- 250 W
- Slots
- 2
- Length
- 267 mm
- PCIe
- Gen 3 x16
- NVLink
- No
- Cooling
- Passive (needs server airflow)
- Software
- CUDA, Vulkan
- Released
- 2016
- Availability
- used only
- Price basis
- Not sold new.
- Maker page
- images.nvidia.com
Sources
- images.nvidia.com: 24 GB, 346 GB/s, PCIe 3.0 x16, 4.4 x 10.5 in dual slot full height, 250 W, 12 TFLOPS FP32, 47 TOPS INT8, server-optimized
- docs.nvidia.com: CUDA 13 release notes: removed support for Maxwell, Pascal and Volta GPUs
- github.com: owners still running Pascal cards on llama.cpp