L40S
48 GiB vendor-published framebuffer. Catalogue open-weight models that fit, ranked by measured quality. NVIDIA does not pick the winner.
48 GiB framebuffer. Source: https://www.nvidia.com/en-gb/data-center/l40s/
7 of 25 profiled open-weight models fit 48 GiB (vram) at Q8_0, 8192 context. Best rated that fits: Gemma 4 31B.
l40s: vendor 48 GiB framebuffer, not free-after-driver.
Among models that fit
Highest LMArena Elo among models that fit: Gemma 4 31B 1443.4. Elo is a human-preference scale, not the intelligence index.
The arena boards are separate scales and are never blended. A missing cell is unmeasured, not zero. A model’s highest axis is relative to its own scores, not a claim it leads the catalogue.
Rated models that fit
| Model | Human preference | Quant that fits | Estimated GiB | Headroom |
|---|---|---|---|---|
Gemma 4 31B Fit assumptions
| 1443.4 | Q8_0 | 33.619 | 14.381 |
Qwen3.8 27B Fit assumptions
| 1440.7 | Q8_0 | 30.188 | 17.813 |
Phi 4 Fit assumptions
| 1216.7 | F16 | 30.563 | 17.438 |
Fits, unrated
- Muse Glimmer 30B — Q8_0, 32.856 GiB unrated
- Ministral 3 14B 2512 — F16, 30.05 GiB unrated
- Ministral 3 8B 2512 — F16, 19.663 GiB unrated
- Ministral 3 3B 2512 — F16, 9.413 GiB unrated
Does not fit, or not profiled
- cohere/command-a-plus — even Q4_K_M wants 132.618 GiB; pool is 48 GiB
- deepseek/deepseek-v4-flash-vision-exp — even Q4_K_M wants 184.902 GiB; pool is 48 GiB
- z-ai/glm-5.3 — even Q4_K_M wants 455.624 GiB; pool is 48 GiB
- qwen/qwen3.8-2.4t-a95b — even Q4_K_M wants 1450.719 GiB; pool is 48 GiB
- moonshotai/kimi-k3 — even Q4_K_M wants 1691.5 GiB; pool is 48 GiB
- z-ai/glm-5.2 — even Q4_K_M wants 455.624 GiB; pool is 48 GiB
- moonshotai/kimi-k2.7-code — even Q4_K_M wants 604.75 GiB; pool is 48 GiB
- nvidia/nemotron-3-ultra-550b-a55b — even Q4_K_M wants 333.063 GiB; pool is 48 GiB
- minimax/minimax-m3 — even Q4_K_M wants 259.405 GiB; pool is 48 GiB
- deepseek/deepseek-v4-pro — even Q4_K_M wants 967 GiB; pool is 48 GiB
- deepseek/deepseek-v4-flash — even Q4_K_M wants 172.465 GiB; pool is 48 GiB
- mistralai/mistral-small-2603 — even Q4_K_M wants 73.2 GiB; pool is 48 GiB
- mistralai/mistral-large-2512 — even Q4_K_M wants 408.531 GiB; pool is 48 GiB
- deepseek/deepseek-v3.2 — even Q4_K_M wants 406.116 GiB; pool is 48 GiB
- openai/gpt-oss-120b — even Q4_K_M wants 72.201 GiB; pool is 48 GiB
- meta-llama/llama-4-maverick — even Q4_K_M wants 242.5 GiB; pool is 48 GiB
- meta-llama/llama-4-scout — even Q4_K_M wants 66.809 GiB; pool is 48 GiB
- cohere/command-a — even Q4_K_M wants 68.016 GiB; pool is 48 GiB
GPU
How the number is made
Weights: a measured GGUF size when we have one, otherwise total parameters × bits per weight ÷ 8. Q4_K_M is treated as 4.83 bits/param. MoE memory uses total parameters, not active parameters.
KV cache, when layer and head geometry is sourced: 2 × kvLayers × kvHeads × headDim × context × 2 bytes (FP16, batch 1). Hybrid models use attention-layer count, not every layer. If geometry is missing, cache is omitted and that is stated.
1 GiB is added for CUDA/runtime. Apple unified memory uses RAM as the pool. CPU-only is a memory fit, not a speed claim. DDR generation is ignored for fit — it changes bandwidth, not whether the weights sit in memory.
RAM type (DDR4/DDR5/LPDDR) is not a field. It does not change whether a model fits.
Questions this page answers
What LLM can I run on a L40S?
7 of 25 profiled open-weight models fit 48 GiB (vram) at Q8_0, 8192 context. Best rated that fits: Gemma 4 31B.
How much memory does a L40S have?
48 GiB vendor-published framebuffer (https://www.nvidia.com/en-gb/data-center/l40s/), not free-after-driver.
Does NVIDIA pick the winner?
No. Ranking is arithmetic on sourced parameter counts. Quality leads. Unrated models that fit are listed separately, never scored zero.