Your rig
L4 · 24 GB
32 GB system RAM · Ada Lovelace · 300 GB/s · 22.4 GB usable VRAM
Which one should I actually run?
Best models for your L4
What do you want to do? Each row is the best-ranked model on that use case's shortlist that fits at Q6_K and 4K. The ordering is editorial; the sizes and speeds are computed. How the shortlists work.
All
General assistantCodingReasoning & mathsWritingVisionAgents & tool useTranslationLow latencyLong context
| Recommendation | Model | Why | Quant · total | Tok/s est. | |
|---|---|---|---|---|---|
| Best overall | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
Best capability per gigabyte: a current-generation dense 27B with vision and 262K context. | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Best for coding | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
Terminal-Bench 73, DeepSWE 42 — a generation ahead of anything else that fits 24 GB. | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Best reasoning | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
Thinking mode plus the best agentic reasoning at 24 GB. | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Best for writing | Gemma 4 12B Google · Apache 2.0 |
The 12 GB writing pick; 140+ languages. | Q6_K · 10.3 GB | 20 | Runs great |
| Best vision | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
Image and video in, OSWorld 84 — the strongest local vision-language model at 24 GB. | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Best for agents | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
Computer use and tool calling are what 3.8 was trained for (OSWorld 84). | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Best translation | Qwen3.8 27B New this week Alibaba · Apache 2.0 |
119 languages, and Qwen translations read as idiomatic rather than literal. | Q6_K · 22.1 GB | 8.6 | Tight fit |
| Fastest good model | Gemma 4 26B-A4B Google · Apache 2.0 |
3.8B active and a small file — the quickest Gemma 4 with vision. | Q6_K · 21.2 GB | 24 | Tight fit |
| Best long context | Gemma 4 12B Google · Apache 2.0 |
256K on a 12 GB card. | Q6_K · 18.1 GB at 128K context |
20 | Runs great |
Everything that fits
Runs great
38 of 76 · fits with headroom for a longer prompt| Model | Params | Quant | Weights | +KV 4K | Total | VRAM headroom | Tok/s | Runtime | |
|---|---|---|---|---|---|---|---|---|---|
| gpt-oss 20B | 20.9B MoE / 3.6B act | MXFP4 | 10.8 GB | 0.10 GB | 11.5 GB |
10.9 GB free
|
37 | Ollama | Run |
| Qwen3 14B | 14.8B | Q6_K | 11.3 GB | 0.63 GB | 12.5 GB |
9.9 GB free
|
16 | Ollama | Run |
| DeepSeek-R1-Distill-Qwen 14B | 14.8B | Q6_K | 11.3 GB | 0.75 GB | 12.7 GB |
9.7 GB free
|
16 | Ollama | Run |
| Qwen2.5-Coder 14B | 14.8B | Q6_K | 11.3 GB | 0.75 GB | 12.7 GB |
9.7 GB free
|
16 | Ollama | Run |
| Phi-4 14B | 14.7B | Q6_K | 11.2 GB | 0.78 GB | 12.6 GB |
9.8 GB free
|
16 | Ollama | Run |
| Ministral 3 14B | 13.9B | Q6_K | 10.6 GB | 0.63 GB | 11.8 GB |
10.6 GB free
|
17 | Ollama | Run |
| Gemma 3 12B | 12.2B | Q6_K | 9.3 GB | 0.56 GB | 10.5 GB |
11.9 GB free
|
19 | Ollama | Run |
| Mistral NeMo 12B | 12.2B | Q6_K | 9.3 GB | 0.63 GB | 10.5 GB |
11.9 GB free
|
19 | Ollama | Run |
| Gemma 4 12B | 12B | Q6_K | 9.2 GB | 0.56 GB | 10.3 GB |
12.1 GB free
|
20 | Ollama | Run |
| Qwen3.5 9B | 9.65B | Q6_K | 7.4 GB | 0.13 GB | 8.1 GB |
14.3 GB free
|
25 | Ollama | Run |
| Ornith 1.5 9B New this week | 9.41B | Q6_K | 7.2 GB | 0.13 GB | 7.9 GB |
14.5 GB free
|
25 | Ollama | Run |
| Ministral 3 8B | 8.92B | Q6_K | 6.8 GB | 0.53 GB | 7.9 GB |
14.5 GB free
|
27 | Ollama | Run |
| Granite 4.1 8B | 8.79B | Q6_K | 6.7 GB | 0.63 GB | 7.9 GB |
14.5 GB free
|
27 | Ollama | Run |
| LFM2.5 8B-A1B | 8.47B MoE / 1.5B act | Q6_K | 6.5 GB | 0.05 GB | 7.1 GB |
15.3 GB free
|
61 | Ollama | Run |
| Fara 7B | 8.29B | Q6_K | 6.3 GB | 0.22 GB | 7.1 GB |
15.3 GB free
|
29 | llama.cpp | Run |
| Qwen3 8B | 8.19B | Q6_K | 6.3 GB | 0.56 GB | 7.4 GB |
15.0 GB free
|
29 | Ollama | Run |
| Llama 3.1 8B Instruct | 8.03B | Q6_K | 6.1 GB | 0.50 GB | 7.2 GB |
15.2 GB free
|
30 | Ollama | Run |
| Gemma 4 E4B | 8.0B | Q6_K | 6.1 GB | 0.09 GB | 6.8 GB |
15.6 GB free
|
30 | Ollama | Run |
| Ling 3.0 Tiny 7.9B-A1.3B New | 7.9B MoE / 1.3B act | Q6_K | 6.0 GB | 0.03 GB | 6.7 GB |
15.7 GB free
|
70 | llama.cpp | Run |
| DeepSeek-R1-Distill-Qwen 7B | 7.62B | Q6_K | 5.8 GB | 0.22 GB | 6.6 GB |
15.8 GB free
|
31 | Ollama | Run |
| Qwen2.5-Coder 7B | 7.62B | Q6_K | 5.8 GB | 0.22 GB | 6.6 GB |
15.8 GB free
|
31 | Ollama | Run |
| Olmo 3 7B Instruct | 7.3B | Q6_K | 5.6 GB | 2.00 GB | 8.2 GB |
14.2 GB free
|
33 | Ollama | Run |
| Mistral 7B Instruct v0.3 | 7.25B | Q6_K | 5.5 GB | 0.50 GB | 6.6 GB |
15.8 GB free
|
33 | Ollama | Run |
| Gemma 4 E2B | 5.1B | Q6_K | 3.9 GB | 0.04 GB | 4.5 GB |
17.9 GB free
|
47 | Ollama | Run |
| Qwen3.5 4B | 4.66B | Q6_K | 3.6 GB | 0.13 GB | 4.3 GB |
18.1 GB free
|
51 | Ollama | Run |
| Gemma 3 4B | 4.3B | Q6_K | 3.3 GB | 0.20 GB | 4.1 GB |
18.3 GB free
|
55 | Ollama | Run |
| Qwen3 4B | 4.02B | Q6_K | 3.1 GB | 0.56 GB | 4.2 GB |
18.2 GB free
|
59 | Ollama | Run |
| Ministral 3 3B | 3.85B | Q6_K | 2.9 GB | 0.41 GB | 3.9 GB |
18.5 GB free
|
62 | Ollama | Run |
| Phi-4-mini 3.8B | 3.84B | Q6_K | 2.9 GB | 0.50 GB | 4.0 GB |
18.4 GB free
|
62 | Ollama | Run |
| Granite 4.1 3B | 3.4B | Q6_K | 2.6 GB | 0.31 GB | 3.5 GB |
18.9 GB free
|
70 | Ollama | Run |
| Llama 3.2 3B Instruct | 3.21B | Q6_K | 2.5 GB | 0.44 GB | 3.5 GB |
18.9 GB free
|
74 | Ollama | Run |
| LFM2.5 2.6B New | 2.7B | Q6_K | 2.1 GB | 0.06 GB | 2.7 GB |
19.7 GB free
|
88 | llama.cpp | Run |
| Qwen3.5 2B | 2.27B | Q6_K | 1.7 GB | 0.05 GB | 2.4 GB |
20.0 GB free
|
105 | Ollama | Run |
| Qwen3 1.7B | 1.72B | Q6_K | 1.3 GB | 0.44 GB | 2.4 GB |
20.0 GB free
|
138 | Ollama | Run |
| Llama 3.2 1B Instruct | 1.24B | Q6_K | 0.9 GB | 0.13 GB | 1.7 GB |
20.7 GB free
|
192 | Ollama | Run |
| Gemma 3 1B | 1.0B | Q6_K | 0.8 GB | 0.03 GB | 1.4 GB |
21.0 GB free
|
238 | Ollama | Run |
| Qwen3.5 0.8B | 0.87B | Q6_K | 0.7 GB | 0.05 GB | 1.3 GB |
21.1 GB free
|
273 | Ollama | Run |
| Qwen3 0.6B | 0.6B | Q6_K | 0.5 GB | 0.44 GB | 1.5 GB |
20.9 GB free
|
396 | Ollama | Run |
Runs — tight
6 models · fits at 4K, drop context or quant to go longer| Model | Params | Quant | Weights | +KV 4K | Total | VRAM headroom | Tok/s | Runtime | |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.8 27B New this week | 27.8B | Q6_K | 21.2 GB | 0.25 GB | 22.1 GB |
0.3 GB free
|
8.6 | llama.cpp | Run |
| Qwen3.6 27B | 27.8B | Q6_K | 21.2 GB | 0.25 GB | 22.1 GB |
0.3 GB free
|
8.6 | llama.cpp | Run |
| Gemma 3 27B | 27.4B | Q6_K | 20.9 GB | 0.72 GB | 22.2 GB |
0.2 GB free
|
8.7 | llama.cpp | Run |
| Gemma 4 26B-A4B | 26.5B MoE / 3.8B act | Q6_K | 20.2 GB | 0.35 GB | 21.2 GB |
1.2 GB free
|
24 | llama.cpp | Run |
| Devstral Small 2 24B | 24B | Q6_K | 18.3 GB | 0.63 GB | 19.6 GB |
2.8 GB free
|
9.9 | llama.cpp | Run |
| Mistral Small 3.2 24B | 23.6B | Q6_K | 18.0 GB | 0.63 GB | 19.2 GB |
3.2 GB free
|
10 | llama.cpp | Run |
Won't fit in VRAM
32 models · some are possible with CPU offload| Model | Params | Quant | Weights | +KV 4K | Total | Over budget | Tok/s | Runtime | |
|---|---|---|---|---|---|---|---|---|---|
| Qwen3.6 35B-A3B | 35.9B MoE / 3.3B act | Q6_K | 27.4 GB | 0.08 GB | 28.1 GB | +5.7 GB | ~15 | llama.cpp | Why |
| Ornith 1.5 35B-A3B New this week | 35.9B MoE / 3.3B act | Q6_K | 27.4 GB | 0.08 GB | 28.1 GB | +5.7 GB | ~15 | llama.cpp | Why |
| LLM-jp 4 33B Thinking New this week | 33.2B | Q6_K | 25.4 GB | 1.00 GB | 27.0 GB | +4.6 GB | ~4.2 | llama.cpp | Why |
| Qwen3 32B | 32.8B | Q6_K | 25.0 GB | 1.00 GB | 26.6 GB | +4.2 GB | ~4.3 | llama.cpp | Why |
| DeepSeek-R1-Distill-Qwen 32B | 32.8B | Q6_K | 25.0 GB | 1.00 GB | 26.6 GB | +4.2 GB | ~4.3 | llama.cpp | Why |
| Qwen2.5-Coder 32B | 32.8B | Q6_K | 25.0 GB | 1.00 GB | 26.6 GB | +4.2 GB | ~4.3 | llama.cpp | Why |
| Olmo 3.1 32B Instruct | 32.2B | Q6_K | 24.6 GB | 1.00 GB | 26.2 GB | +3.8 GB | ~4.6 | llama.cpp | Why |
| Nemotron 3.5 Lightning 30B-A3B New | 31.6B MoE / 3.2B act | Q6_K | 24.1 GB | 0.02 GB | 24.8 GB | +2.4 GB | ~21 | llama.cpp | Why |
| Gemma 4 31B | 31.3B | Q6_K | 23.9 GB | 1.41 GB | 25.9 GB | +3.5 GB | ~4.8 | llama.cpp | Why |
| GLM-4.7-Flash 30B-A3B | 31.2B MoE / 3B act | Q6_K | 23.8 GB | 0.21 GB | 24.6 GB | +2.2 GB | ~22 | llama.cpp | Why |
| Qwen3 30B-A3B | 30.5B MoE / 3.3B act | Q6_K | 23.3 GB | 0.38 GB | 24.3 GB | +1.9 GB | ~21 | llama.cpp | Why |
| Qwen3 Coder 30B-A3B | 30.5B MoE / 3.3B act | Q6_K | 23.3 GB | 0.38 GB | 24.3 GB | +1.9 GB | ~21 | llama.cpp | Why |
| Muse Glimmer 30B New | 29.8B | Q6_K | 22.8 GB | 0.13 GB | 23.5 GB | +1.1 GB | ~6.7 | llama.cpp | Why |
| Granite 4.1 30B | 28.9B | Q6_K | 22.1 GB | 1.00 GB | 23.7 GB | +1.3 GB | ~6.7 | llama.cpp | Why |
| Kimi K3 2.8T-A104B New | 2780B MoE / 104B act | Q6_K | 2123.0 GB | 0.11 GB | 2123.7 GB | +2101.3 GB | ~0.2 | llama.cpp | Why |
| Qwen3.8 2.4T-A95B New | 2446B MoE / 95B act | Q6_K | 1868.0 GB | 0.36 GB | 1868.9 GB | +1846.5 GB | ~0.2 | llama.cpp | Why |
| DeepSeek V4 Pro 1.6T-A49B New | 1650B MoE / 49B act | Q6_K | 1260.1 GB | 0.27 GB | 1260.9 GB | +1238.5 GB | ~0.4 | llama.cpp | Why |
| Kimi K2.6 1T-A32B | 1027B MoE / 32B act | Q6_K | 784.3 GB | 0.27 GB | 785.2 GB | +762.8 GB | ~0.6 | llama.cpp | Why |
| GLM-5.2 744B-A40B | 753B MoE / 40B act | Q6_K | 575.1 GB | 0.34 GB | 576.0 GB | +553.6 GB | ~0.5 | llama.cpp | Why |
| DeepSeek-R1 671B | 671B MoE / 37B act | Q6_K | 512.4 GB | 0.27 GB | 513.3 GB | +490.9 GB | ~0.5 | llama.cpp | Why |
| Qwen3.5 397B-A17B | 403B MoE / 17B act | Q6_K | 307.8 GB | 0.12 GB | 308.5 GB | +286.1 GB | ~1.1 | llama.cpp | Why |
| Ornith 1.5 397B-A17B New this week | 397B MoE / 17B act | Q6_K | 303.2 GB | 0.12 GB | 303.9 GB | +281.5 GB | ~1.1 | llama.cpp | Why |
| DeepSeek V4 Flash 284B-A13B New | 284B MoE / 13B act | Q6_K | 216.9 GB | 0.19 GB | 217.7 GB | +195.3 GB | ~1.5 | llama.cpp | Why |
| Qwen3 235B-A22B | 235B MoE / 22B act | Q6_K | 179.5 GB | 0.73 GB | 180.8 GB | +158.4 GB | ~0.9 | llama.cpp | Why |
| Qwen3.5 122B-A10B | 125B MoE / 10B act | Q6_K | 95.5 GB | 0.09 GB | 96.2 GB | +73.8 GB | ~2.2 | llama.cpp | Why |
| Nemotron 3 Super 120B-A12B | 124B MoE / 12B act | Q6_K | 94.7 GB | 0.03 GB | 95.3 GB | +72.9 GB | ~1.9 | llama.cpp | Why |
| Ling 3.0 Flash 124B-A5B New | 124B MoE / 5.1B act | Q6_K | 94.7 GB | 0.03 GB | 95.3 GB | +72.9 GB | ~4.4 | llama.cpp | Why |
| Mistral Small 4 119B-A6B | 119B MoE / 6.5B act | Q6_K | 90.9 GB | 0.09 GB | 91.6 GB | +69.2 GB | ~3.5 | llama.cpp | Why |
| gpt-oss 120B | 117B MoE / 5.1B act | MXFP4 | 60.5 GB | 0.15 GB | 61.3 GB | +38.9 GB | ~7.4 | llama.cpp | Why |
| Llama 4 Scout 109B-A17B | 109B MoE / 17B act | Q6_K | 83.2 GB | 0.75 GB | 84.6 GB | +62.2 GB | ~1.3 | llama.cpp | Why |
| Llama 3.3 70B Instruct | 70.6B | Q6_K | 53.9 GB | 1.25 GB | 55.8 GB | +33.4 GB | ~1.0 | llama.cpp | Why |
| DeepSeek-R1-Distill-Llama 70B | 70.6B | Q6_K | 53.9 GB | 1.25 GB | 55.8 GB | +33.4 GB | ~1.0 | llama.cpp | Why |
Totals = quantised weights + f16 KV cache at 4K tokens + 0.6 GB runtime overhead, against 22.4 GB usable VRAM (24 GB less display and driver reserve). Tokens per second are estimated from 300 GB/s peak memory bandwidth at batch 1 — not measured. Read the method.