What can NVIDIA L4 run for agents?
Build: NVIDIA L4 + — + 32 GB RAM (windows)
Runs comfortably81 models
Ranked by fit for agents use case + predicted speed. Click a row for VRAM breakdown.
Quant: Q8_0Context: 8,192VRAM: 14.7 GBHeadroom: 9.3 GBollama run hermes3:8b115tok/sEstimated
ollama run hermes3:8bQuant: Q5_K_MContext: 8,192VRAM: 19.8 GBHeadroom: 4.2 GBollama run qwen2.5:14b101tok/sEstimated
ollama run qwen2.5:14bQuant: Q4_K_MContext: 2,048VRAM: 19.5 GBHeadroom: 4.5 GBollama run dolphin-mistral:24b67tok/sEstimated
ollama run dolphin-mistral:24bQuant: Q8_0Context: 8,192VRAM: 10.8 GBHeadroom: 13.2 GBollama run qwen2.5:7b131tok/sEstimated
ollama run qwen2.5:7bQuant: Q4_K_MContext: 8,192VRAM: 11.0 GBHeadroom: 13.0 GB202tok/sEstimated
Quant: Q5_K_MContext: 8,192VRAM: 10.7 GBHeadroom: 13.3 GBollama run mistral:7b203tok/sEstimated
ollama run mistral:7bQuant: Q4_K_MContext: 2,048VRAM: 19.5 GBHeadroom: 4.5 GBollama run mistral-small:24b67tok/sEstimated
ollama run mistral-small:24bQuant: FP16Context: 8,192VRAM: 19.8 GBHeadroom: 4.2 GBollama run llama3.1:8b61tok/sEstimated
ollama run llama3.1:8bQuant: Q4_K_MContext: 8,192VRAM: 17.6 GBHeadroom: 6.4 GBollama run qwen3:14b115tok/sEstimated
ollama run qwen3:14bQuant: Q4_K_MContext: 8,192VRAM: 17.6 GBHeadroom: 6.4 GB115tok/sEstimated
Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 14.1 GB231tok/sEstimated
Quant: Q4_K_MContext: 8,192VRAM: 9.9 GBHeadroom: 14.1 GB231tok/sEstimated
Runs with tradeoffs32 models
Tight VRAM, partial CPU offload, or context-limited.
Quant: Q4_K_MContext: 8,192VRAM: 35.7 GBHeadroom: 7.5 GB- • Partial CPU offload: ~33% of layers run on CPU
ollama run qwen3:30b54tok/sEstimated
- • Partial CPU offload: ~33% of layers run on CPU
ollama run qwen3:30bQuant: Q4_K_MContext: 2,048VRAM: 22.5 GBHeadroom: 1.5 GB- • Tight VRAM fit — only 1.5 GB headroom left for context growth
ollama run qwen3.8:27b60tok/sEstimated
- • Tight VRAM fit — only 1.5 GB headroom left for context growth
ollama run qwen3.8:27bQuant: Q4_K_MContext: 8,192VRAM: 23.9 GBHeadroom: 0.1 GB- • Tight VRAM fit — only 0.1 GB headroom left for context growth
ollama run qwen2.5-coder:32b50tok/sEstimated
- • Tight VRAM fit — only 0.1 GB headroom left for context growth
ollama run qwen2.5-coder:32bQuant: Q4_K_MContext: 8,192VRAM: 37.8 GBHeadroom: 5.4 GB- • Partial CPU offload: ~36% of layers run on CPU
ollama run qwen3:32b50tok/sEstimated
- • Partial CPU offload: ~36% of layers run on CPU
ollama run qwen3:32bQuant: Q4_K_MContext: 8,192VRAM: 37.8 GBHeadroom: 5.4 GB- • Partial CPU offload: ~36% of layers run on CPU
ollama run qwen2.5:32b50tok/sEstimated
- • Partial CPU offload: ~36% of layers run on CPU
ollama run qwen2.5:32bQuant: Q4_K_MContext: 8,192VRAM: 35.7 GBHeadroom: 7.5 GB- • Partial CPU offload: ~33% of layers run on CPU
ollama run nemotron3:nano54tok/sEstimated
- • Partial CPU offload: ~33% of layers run on CPU
ollama run nemotron3:nanoQuant: Q4_K_MContext: 2,048VRAM: 37.1 GBHeadroom: 6.1 GB- • Partial CPU offload: ~35% of layers run on CPU
ollama run mixtral:8x7b34tok/sEstimated
- • Partial CPU offload: ~35% of layers run on CPU
ollama run mixtral:8x7bQuant: Q4_K_MContext: 2,048VRAM: 21.9 GBHeadroom: 2.1 GB- • Tight VRAM fit — only 2.1 GB headroom left for context growth
ollama run gemma4:26b-moe62tok/sEstimated
- • Tight VRAM fit — only 2.1 GB headroom left for context growth
ollama run gemma4:26b-moeWhat if you upgraded?
Hypothetical scenarios. We re-ran the compatibility engine for each.
+32 GB system RAM
Check the current price
Adds 32 GB to your CPU-offload working set. Helps when models don't quite fit in VRAM.
Unlocks: 49 new comfortable, 48 new tradeoff
- • Qwen 3 0.6B
- • Qwen 3 1.7B
- • Llama 3.2 3B Instruct
- • Gemma 3 270M
Upgrade to NVIDIA GeForce RTX 5090
Launch MSRP $1,999 (2025). Check the current price.
32 GB VRAM (vs your 24 GB) plus a bandwidth jump from ~? GB/s to ~1792 GB/s.
Unlocks: 73 new comfortable
- • Qwen 3 0.6B
- • Qwen 3 30B-A3B
- • Qwen 2.5 Coder 32B Instruct
- • Qwen 3 32B
Add a second NVIDIA L4
Launch MSRP $2,500 (2023). Check the current price.
Tensor parallelism splits the model across both cards, effectively doubling VRAM. Bandwidth doesn't double — runs ~1.5× the single-card speed in practice.
Unlocks: 81 new comfortable
- • Qwen 3 0.6B
- • Qwen 3 30B-A3B
- • Qwen 2.5 Coder 32B Instruct
- • Qwen 3 32B
Some links above are affiliate links. We may earn a commission at no extra cost to you. How we make money.
Won't runtop 5 popular models
Need more memory than you have. Shown for orientation.
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
—
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
—
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
—
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
—
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
—
Even with CPU offload, needs more memory than your VRAM (24 GB) + 60% of system RAM (19 GB) combined.
How to read these numbers
Want a specific benchmark we don't have? Email Contact support and we'll prioritize it.