# NVIDIA Tesla P40 for local LLMs

> The cheapest 24GB card, if you can cool it and accept Pascal's slow compute.

The Tesla P40 has 24GB of GDDR5 and 346 GB/s of memory bandwidth, which decides how big a model fits and how fast it generates text. It comfortably runs models up to about 35B parameters: the largest in our catalog is Qwen3.6 35B-A3B at Q3_K_M (19GB with 8K context, ~58 tokens/sec). Our top all-round pick for it is Qwen3.8 27B (Q4_K_M, ~11 tokens/sec).

## Key specs

- Memory: 24GB GDDR5
- Memory bandwidth: 346 GB/s
- Architecture: Pascal (2016)
- Power: 250W
- Typical price: ~$230 used (checked 2026-10)
- Software: CUDA, Vulkan

## Best models for this GPU

- **Best overall (also coding, reasoning, vision):** Qwen3.8 27B at Q4_K_M, ~18.1GB with 8K context, ~11 tokens/sec
- **Runner-up:** Gemma 4 26B A4B at Q5_K_M, ~20.1GB with 8K context, ~36 tokens/sec
- **Fastest good model (also largest that fits):** Qwen3.6 35B-A3B at Q3_K_M, ~19.2GB with 8K context, ~58 tokens/sec

## Compatibility (best quantization that fits, 8K context)

| Model | Params | Quant | Memory | Speed | Verdict |
| --- | --- | --- | --- | --- | --- |
| Gemma 4 E4B | 8B | Q8_0 | 10.5GB | ~20 tok/s | Runs well |
| Qwen3.5 9B | 9B | Q8_0 | 12.0GB | ~17 tok/s | Runs well |
| Phi-4 | 14.7B | Q8_0 | 18.3GB | ~12 tok/s | Runs well |
| gpt-oss-20b | 20.9B (3.6B active) | MXFP4 | 15.0GB | ~41 tok/s | Runs well |
| Gemma 4 26B A4B | 26B (3.8B active) | Q5_K_M | 20.1GB | ~36 tok/s | Runs well |
| Qwen3.8 27B | 27B | Q4_K_M | 18.1GB | ~11 tok/s | Runs well |
| Gemma 4 31B | 31B | Q4_K_M | 22.5GB | ~9.6 tok/s | Fits, tight |
| Qwen3.6 35B-A3B | 35B (3B active) | Q3_K_M | 19.2GB | ~58 tok/s | Runs well |
| GLM-4.5-Air | 106B (12B active) | Q4_K_M | 68.3GB | ~7.4 tok/s | Partial offload |
| Llama 4 Scout | 109B (17B active) | Q4_K_M | 69.5GB | ~5.3 tok/s | Partial offload |
| gpt-oss-120b | 117B (5.1B active) | MXFP4 | 68.3GB | ~18 tok/s | Partial offload |
| Qwen3.5 122B-A10B | 122B (10B active) | - | - | - | Won't fit |
| DeepSeek V4 Flash | 284B (13B active) | - | - | - | Won't fit |
| GLM-5.3-Flash | 320B (18B active) | - | - | - | Won't fit |
| GLM-5.3 | 744B (40B active) | - | - | - | Won't fit |

## Pros
- 24GB for the lowest price of any card
- Works with llama.cpp and other CUDA 12 builds
- Easy to buy in quantity

## Cons
- Slow FP16 and no tensor cores
- Needs added cooling and an EPS power adapter
- Dropped from CUDA 13

Full page: https://llamabuilds.ai/gpus/tesla-p40 · Best models: https://llamabuilds.ai/gpus/tesla-p40/best-models

Sources: [NVIDIA Tesla P40 product brief (PDF)](https://images.nvidia.com/content/pdf/tesla/Tesla-P40-Product-Brief.pdf), [Flopper.io Tesla P40 spec sheet](https://flopper.io/gpu/nvidia-tesla-p40-24gb/spec-sheet)

---
Source: https://llamabuilds.ai/gpus/tesla-p40 · Updated October 14, 2026 · Llama Builds. Speeds marked ~ are estimates from our bandwidth model (https://llamabuilds.ai/methodology); measured results are labeled as such.