NVIDIA · Pascal · 16 GB VRAM tier
Best AI Models for NVIDIA Tesla P100 PCIe 16GB
Runs 14B models well; 27–31B only just fits.
Pascal datacenter card with full-rate FP16 over HBM2, but no Tensor Cores and only 16 GB. Passive server card; needs added cooling.
| # | Model | Speed | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 146 t/s | 146 tok/s | Q4_K_M · 3.4 GB | |
| 3 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 148 t/s | 148 tok/s | Q4_K_M · 13.1 GB | |
| 4 | Phi 4Microsoft Q4_K_M · 12.5 GB · 50.0 t/s | 50.0 tok/s | Q4_K_M · 12.5 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 57.8 t/s | 57.8 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed | Coding Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | GPT OSS 20BOpenAIHow to run → Q4_K_M · 13.1 GB · 148 t/s | 148 tok/s | Q4_K_M · 13.1 GB | |
| 2 | Q4_K_M · 7.9 GB · 89.8 t/s | 89.8 tok/s | Q4_K_M · 7.9 GB | |
| 3 | Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | overall | Q4_K_M · 6.6 GB |
| 4 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 57.8 t/s | 57.8 tok/s | Q4_K_M · 9.0 GB | |
| 5 | Ling 3.0 TinyInclusion AI Q4_K_M · 5.5 GB · 304 t/s | 304 tok/s | Q4_K_M · 5.5 GB |
| # | Model | Speed | Reasoning Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | Q4_K_M · 6.6 GB | |
| 2 | GPT OSS 20BOpenAI Q4_K_M · 13.1 GB · 148 t/s | 148 tok/s | Q4_K_M · 13.1 GB | |
| 3 | Qwen3 8BAlibaba Q4_K_M · 7.6 GB · 86.2 t/s | 86.2 tok/s | Q4_K_M · 7.6 GB | |
| 4 | Q4_K_M · 7.9 GB · 89.8 t/s | 89.8 tok/s | Q4_K_M · 7.9 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 57.8 t/s | 57.8 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed | Long context Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 146 t/s | 146 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Mistral Nemo Instruct 2407Mistral AI Q4_K_M · 11.0 GB · 59.0 t/s | 59.0 tok/s | Q4_K_M · 11.0 GB | |
| 4 | Q4_K_M · 7.9 GB · 89.8 t/s | 89.8 tok/s | Q4_K_M · 7.9 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 57.8 t/s | 57.8 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed ↓ | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | GPT OSS 20BOpenAIHow to run → Q4_K_M · 13.1 GB · 148 t/s | 148 tok/s | Q4_K_M · 13.1 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 146 t/s | 146 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | Q4_K_M · 6.6 GB | |
| 4 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 57.8 t/s | 57.8 tok/s | Q4_K_M · 9.0 GB | |
| 5 | Phi 4Microsoft Q4_K_M · 12.5 GB · 50.0 t/s | 50.0 tok/s | Q4_K_M · 12.5 GB |
| # | Model | Speed | Vision Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 74.8 t/s | 74.8 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 146 t/s | 146 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Gemma 3 12B ITGoogle Q4_K_M · 10.9 GB · 59.2 t/s | 59.2 tok/s | Q4_K_M · 10.9 GB | |
| 4 | Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 168 t/s | 168 tok/s | Q4_K_M · 5.0 GB | |
| 5 | Q4_K_M · 6.5 GB · 76.6 t/s | 76.6 tok/s | Q4_K_M · 6.5 GB |
The Tesla P100 PCIe 16GB runs 14B models well. Pascal generation, released 2016.
Tesla P100 PCIe 16GB strengths and weaknesses
Strengths
- Runs 35B MoE models: Qwen3.6 35B A3B at 153 tok/s in 15.7 GB
Weaknesses
- 70B models don't fit: Llama 3.3 70B Instruct needs 51.3 GB at Q4 vs 16 GB here
- Slow prompt reading: A 4K-token prompt on Qwen3.5 9B waits 9.4 s for the first token
- No FP8 or NVFP4: Pascal predates both; use GGUF, AWQ or GPTQ builds
Output speed depends on memory bandwidth (732 GB/s). Grey rows need more than 16 GB.
- 9BQwen3.5 9B · Q4_K_M · 6.6 GB74.8tok/s
- 14BQwen3 14B · Q4_K_M · 11.9 GB50.1tok/s
- 31.7tok/s
- 153tok/s
- 70BLlama 3.3 70B Instruct · not at 4-bitneeds 51.3 GB
- 120B MoEGPT OSS 120B · not at 4-bitneeds 70.8 GB
Estimated speeds: Q4_K_M, 4K context, one user. How we estimate
All 1057 models on the Tesla P100 PCIe 16GB
Featured onlyBest quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works
| Model | Quant | VRAM | Speed | Context | Score | Fit |
|---|---|---|---|---|---|---|
| Q4_K_M | 2.8 GB | 424.8 t/s | — | S100, Runs great | ||
| Q4_K_M | 4.5 GB | 151.5 t/s | 262K | S97, Runs great | ||
Qwopus3.5 4B CoderChatReasoningFunctionsCode | BF16 | 10.0 GB | 48.6 t/s | 262K | A79, Runs well | |
DeepHat V1 7BChatCode | Q4_K_M | 5.8 GB | 95.4 t/s | 33K | S90, Runs great | |
| Q4_K_M | 7.6 GB | 86.2 t/s | 41K | S88, Runs great | ||
SmolLM3 3B ONNXChat | Q4_K_M | 3.3 GB | 211.5 t/s | 66K | S100, Runs great | |
Dots.ocrVision | Q4_K_M | 2.6 GB | 218.3 t/s | 131K | S100, Runs great | |
| Q4_K_M | 3.7 GB | 179.5 t/s | 8K | S99, Runs great | ||
| Q4_K_M | 0.0 GB | 47580.0 t/s | 1K | S100, Runs great | ||
Translategemma 4B ITVision | Q4_K_M | 5 GB | 167.5 t/s | — | S98, Runs great | |
North Mini Code 1.0ChatCodeFunctions | Q3_K_M | 16.0 GB | 162.0 t/s | 500K | C44, Usable | |
| Q4_K_M | 0.3 GB | 1586.0 t/s | 2K | S100, Runs great | ||
| Q4_K_M | 0.5 GB | 1586.0 t/s | 33K | S100, Runs great | ||
| Q4_K_M | 4.2 GB | 164.1 t/s | 33K | S97, Runs great | ||
Qwen3 Embedding 0.6BChatFunctions | Q4_K_M | 1.6 GB | 617.9 t/s | 33K | S100, Runs great |
Specs
- Architecture
- Pascal
- VRAM
- 16 GB HBM2
- Memory bandwidth
- 732.0 GB/s
- FP16 (dense)
- 18.70 TFLOPS
- CUDA cores
- 3,584
- Tensor cores
- 0
- Compute capability
- 6.0
- TDP
- 250 W
- Released
- 2016-06-20
Tesla P100 PCIe 16GB FAQ
What is the largest model the Tesla P100 PCIe 16GB can run?
Gemma 4 26B A4B IT (25.8B) is the largest at 4-bit or better: at Q4_K_S it uses 15.5 GB of 16 GB, and runs at about 141 tok/s. At 3-bit, a bigger model fits too: Gemma 2 27B IT (27.2B) at IQ3_M, 15.0 GB.
Can the Tesla P100 PCIe 16GB run Qwen3.5 9B?
Yes. Qwen3.5 9B fits at Q4_K_M, using 6.6 GB of 16 GB, and it is the best model that runs well on this card.
Can the Tesla P100 PCIe 16GB run 70B models?
No. Llama 3.3 70B Instruct needs 51.3 GB at Q4, more than the 16 GB on this card.
How many tokens per second does the Tesla P100 PCIe 16GB give?
About 74.8 tok/s on Qwen3.5 9B at Q4_K_M, with the first token after 9.4 s on a 4K-token prompt.
What power supply does the Tesla P100 PCIe 16GB need?
Plan a 500 W PSU (2× the 250 W TDP).
Is 16 GB enough for coding and agents?
Yes for coding: GPT OSS 20B, the best of the 2 coding models that fit, runs at about 148 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.