The best open LLMs you can run on 8 GB of VRAM

Every open LLM in our catalog whose weights plus KV-cache actually fit in 8 GB of VRAM at its default quant — the size of an RTX 3060/4060 or an 8 GB laptop GPU. Ranked by how much model you get for that budget (largest parameter count that still fits), with the honest $0-on-your-own-hardware cost for each. You pick the one whose quality you like; we just guarantee it fits.

How this is ranked: Pure objective filter: 'best' = 'fits your 8 GB card.' VRAM is computed by our shared cost engine from params, quant and context — not a quality opinion. We order by model size (most capability per GB) and never claim a #1 is 'smartest.' Quality judgment is the user's.

1. internlm3 8b instruct — internlm, 8.8B · ~8.0 GB VRAM · $0.17/1M API est. · commercial OK
2. Nemotron Labs Diffusion 8B Base — nvidia, 8.5B · ~7.0 GB VRAM · $0.17/1M API est. · non-commercial
3. LFM2.5 8B A1B — LiquidAI, 8.5B · ~7.0 GB VRAM · $0.17/1M API est. · non-commercial
4. granite 3.0 8b instruct — ibm-granite, 8.2B · ~7.0 GB VRAM · $0.17/1M API est. · commercial OK
5. Llama 3.1 8B Instruct — Meta, 8B · ~8.0 GB VRAM · $0.03/1M API · commercial OK
6. Qwen2-VL 7B Instruct — Alibaba, 8B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
7. Llama 3.1 8B Instruct (Abliterated) — mlabonne (community), 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
8. Hermes 3 — Llama 3.1 8B — Nous Research, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
9. Dolphin 3.0 — Llama 3.1 8B — Cognitive Computations, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
10. LLaDA 8B Instruct — GSAI-ML, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
11. DeepSeek R1 Distill Llama 8B — deepseek-ai, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
12. saiga llama3 8b — IlyaGusev, 8B · ~7.0 GB VRAM · $0.16/1M API est. · non-commercial
13. Meta Llama 3.1 8B Instruct — unsloth, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
14. LLaDA 1.5 — GSAI-ML, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
15. Meta Llama 3 8B Instruct — NousResearch, 8B · ~7.0 GB VRAM · $0.16/1M API est. · non-commercial
16. Meta Llama 3.1 8B Instruct — NousResearch, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
17. Llama 3.1 8B Instruct — unsloth, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
18. LLaDA 8B Base — GSAI-ML, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
19. Humanish Roleplay Llama 3.1 8B — vicgalle, 8B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
20. llava onevision qwen2 7b ov — lmms-lab, 8B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
21. Llama 3.1 Nemotron Safety Guard 8B v3 — nvidia, 8B · ~8.0 GB VRAM · $0.16/1M API est. · non-commercial
22. NeuralDaredevil 8B abliterated — mlabonne, 8B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
23. L3 8B Stheno v3.2 — Sao10K, 8B · ~7.0 GB VRAM · $0.16/1M API est. · non-commercial
24. EXAONE 3.5 7.8B Instruct — LGAI-EXAONE, 7.8B · ~8.0 GB VRAM · $0.16/1M API est. · non-commercial
25. hf moshiko — kmhf, 7.8B · ~8.0 GB VRAM · $0.16/1M API est. · non-commercial
26. internlm2 5 7b chat — internlm, 7.7B · ~8.0 GB VRAM · $0.16/1M API est. · non-commercial
27. internlm2 chat 7b — internlm, 7.7B · ~8.0 GB VRAM · $0.16/1M API est. · non-commercial
28. Qwen2 7B Instruct — Qwen, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
29. Qwen2.5 7B — Qwen, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
30. DeepSeek R1 Distill Qwen 7B — deepseek-ai, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
31. Dream v0 Instruct 7B — Dream-org, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
32. Qwen2.5 Coder 7B — Qwen, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
33. Phi mini MoE instruct — microsoft, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
34. Qwen2.5 Math 7B Instruct — Qwen, 7.6B · ~6.0 GB VRAM · $0.16/1M API est. · commercial OK
35. Qwen2.5 7B Instruct — unsloth, 7.6B · ~7.0 GB VRAM · $0.16/1M API est. · commercial OK
36. OLMo 2 1124 7B Instruct — allenai, 7.3B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
37. Mistral 7B Instruct v0.3 — Mistral AI, 7.2B · ~8.0 GB VRAM · $0.20/1M API · commercial OK
38. Mistral 7B Instruct v0.2 — mistralai, 7.2B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
39. Mistral 7B v0.1 — mistralai, 7.2B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK
40. Mistral 7B Instruct v0.1 — mistralai, 7.2B · ~8.0 GB VRAM · $0.16/1M API est. · commercial OK

Showing the top 40 of 173. See all →

More: all "best" lists · cost calculator · all models