_03.png){kind=link}
Best local AI models for an RTX 4060 8 GB
An NVIDIA GeForce RTX 4060 8 GB card has about 7.2 GB of VRAM free for a model after the driver and context cache. These are the most downloaded open language and vision-language models that fit, for Ollama, llama.cpp, LM Studio or vLLM.
8 GB VRAM · about 7.2 GB usable for a model
Buy a RTX 4060 on Amazon →
As an Amazon Associate we earn from qualifying purchases.
Last updated 3 November 2026 · refreshed after every Hugging Face sync · 15 models
| # | Model | Size | Memory | License | Downloads (30 days) | Likes | Est. speed (Q4) |
|---|---|---|---|---|---|---|---|
| 1 | Qwen2.5-1.5B-Instruct Qwen | 1.5B · FP16 | ~4 GB | Apache 2.0 | 4.6M | 928 | ~164 tok/s |
| 2 | Ornith-1.5-9B-GGUF ornith-ai | 9B · GGUF | ~6 GB | MIT | 5.5M | 626 | – |
| 3 | Qwen3-1.7B Qwen | 1.7B · FP16 | ~5 GB | Apache 2.0 | 4.3M | 612 | – |
| 4 | Llama-3.2-1B-Instruct meta-llama | 1B · FP16 | ~3 GB | Llama 3.2 license | 1.8M | 1.9K | ~185 tok/s |
| 5 | Ornith-1.0-9B-GGUF ornith-ai | 9B · GGUF | ~6 GB | MIT | 2.4M | 699 | – |
| 6 | MiniCPM5-2B openbmb | 2B · FP16 | ~5 GB | Apache 2.0 | 1.1M | 1.8K | – |
| 7 | TinyLlama-1.1B-Chat-v1.0 TinyLlama | 1.1B · FP16 | ~3 GB | Apache 2.0 | 944.2K | 1.9K | – |
| 8 | DeepSeek-R1-Distill-Qwen-1.5B deepseek-ai | 1.5B · FP16 | ~4 GB | MIT | 784.6K | 1.7K | – |
| 9 | gemma-3-1b-it | 1B · FP16 | ~3 GB | Gemma license | 826.6K | 1.3K | ~204 tok/s |
| 10 | Llama-3.2-1B meta-llama | 1B · FP16 | ~3 GB | Llama 3.2 license | 479.7K | 2.7K | ~185 tok/s |
| 11 | Qwen3-8B-AWQ Qwen | 8B · AWQ | ~5 GB | Apache 2.0 | 2M | 70 | ~40 tok/s |
| 12 | Qwen2.5-7B-Instruct-AWQ Qwen | 7B · AWQ | ~5 GB | Apache 2.0 | 1.8M | 61 | ~47 tok/s |
| 13 | Qwythos-9B-Claude-Mythos-5-1M-GGUF empero-ai | 9B · GGUF | ~6 GB | Apache 2.0 | 234.9K | 2.9K | – |
| 14 | Sharp-Spark-X2.5-4B-GGUF peculiar-ragdoll | 4B · GGUF | ~3 GB | Apache 2.0 | 838.9K | 154 | – |
| 15 | gemma-7b | 7B · GGUF | ~5 GB | Gemma license | 136.7K | 3.5K | – |
How we rank
Language and vision-language models in GGUF, FP16, AWQ, GPTQ format whose published weights need about 2–7.2 GB. We include models seen on Hugging Face in the last 14 days with at least 10,000 downloads in the last 30 days. Score = log10(30-day downloads) + ½ × log10(likes + 1), so popularity leads and likes break near-ties. We keep one entry per model name and leave out adult and uncensored builds. Memory is estimated from parameter count and the published file format (16-bit ≈ 2 bytes per parameter, 4-bit ≈ 0.5) plus 20% headroom; long contexts need more. A quantised copy of a bigger model may also fit — check the model page.
Speed data: ModelFit (CC BY 4.0), from its open dataset, refreshed weekly. These are modelled estimates of decode speed for a 4-bit (Q4) build of the same model at a 4K context, not measured benchmarks, and they are not an endorsement by ModelFit. A dash means ModelFit has no estimate for that model.