NVIDIA RTX 4060
The NVIDIA RTX 4060 has 8 GB VRAM and 272 GB/s memory bandwidth. It can run 23 of our 80 tracked models natively in VRAM at 8k context.
With 8 GB GDDR6, the NVIDIA RTX 4060 is a consumer-tier GPU that can run 23 models natively. It's best for smaller models under 8B parameters.
The NVIDIA RTX 4060 is the entry-level Ada Lovelace consumer GPU with 8GB GDDR6 at 272 GB/s. At 8GB VRAM, it is limited to small models — 7B at Q4_K_M fits tightly, and anything larger requires CPU offload. Best suited for users testing small LLMs or running compact coding assistants on a budget.
NVIDIA RTX 4060: 8GB GDDR6 at 272 GB/s — entry-level Ada.
7B at Q4-Q5 native. 8B models tight. ~6-10 t/s for 7B.
Full CUDA support. 8GB is the absolute minimum for practical LLM use.
| Vendor | NVIDIA |
| Architecture | Ada Lovelace |
| VRAM | 8 GB |
| Memory type | GDDR6 |
| Memory bandwidth | 272 GB/s |
| Compute backend | CUDA |
| Tier | Consumer |
| Released | 2023 |
| Models (native) | 23 / 80 |
| Models (offload) | 25 / 80 |
Software: Full llama.cpp and Ollama support out of the box. CUDA 12.x recommended; driver ≥ 525 required.
Popular models for this GPU
Models this GPU runs natively in VRAM (23)
- Bonsai 27B27B · MMLU-Pro 81.51-bit (Q1_0) · ~32.7 t/s
- Phi-4 14B Instruct14B · MMLU-Pro 70.4Q2_K · ~26.5 t/s
- Mistral Nemo 12B Instruct12.2B · MMLU-Pro 35.6Q2_K · ~29.5 t/s
- Gemma 3 12B Instruct12.2B · MMLU-Pro 60.6Q2_K · ~31 t/s
- Gemma 2 9B Instruct9.2B · MMLU-Pro 32.0Q2_K · ~28 t/s
- Llama 3.1 8B Instruct8B · MMLU-Pro 48.3NVFP4 · ~34.8 t/s
- DeepSeek R1 Distill Llama 8B8B · MMLU-Pro 41.0NVFP4 · ~34.8 t/s
- Qwen3 8B8B · MMLU-Pro 56.7NVFP4 · ~33.9 t/s
- Qwen 2.5 7B Instruct7.6B · MMLU-Pro 56.3NVFP4 · ~41.4 t/s
- Mistral 7B Instruct v0.37.25B · MMLU-Pro 30.0NVFP4 · ~37.6 t/s
- Gemma 3 4B Instruct4B · MMLU-Pro 43.6NVFP4 · ~70.6 t/s
- Gemma 4 E4B4B · MMLU-Pro 69.4NVFP4 · ~58.8 t/s
- Phi-3.5 Mini Instruct3.8B · MMLU-Pro 47.4NVFP4 · ~34.5 t/s
- Llama 3.2 3B Instruct3.2B · MMLU-Pro 24.0NVFP4 · ~69.6 t/s
- Qwen 2.5 3B Instruct3.1B · MMLU-Pro 32.4BF16 · ~27.2 t/s
- Gemma 2 2B Instruct2.6B · MMLU-Pro 17.8BF16 · ~29.1 t/s
- Gemma 4 E2B2B · MMLU-Pro 60.0BF16 · ~40.2 t/s
- SmolLM2 1.7B Instruct1.7B · MMLU-Pro 19.0BF16 · ~35.3 t/s
- Qwen 2.5 1.5B Instruct1.5B · MMLU-Pro 16.8FP32 · ~28.4 t/s
- Llama 3.2 1B Instruct1.24B · MMLU-Pro 12.5FP32 · ~33.8 t/s
- Gemma 3 1B Instruct1B · MMLU-Pro 14.7FP32 · ~40.9 t/s
- Qwen 2.5 0.5B Instruct0.5B · MMLU-Pro 10.0FP32 · ~84.2 t/s
- SmolLM2 360M Instruct0.36B · MMLU-Pro 8.0FP32 · ~99.6 t/s
Models that fit with CPU offload (25)
These use system RAM for layers that don't fit in VRAM — expect much slower inference.
- Llama 3.3 70B Instruct70B · MMLU-Pro 68.9Q2_K · ~1.1 t/s
- DeepSeek R1 Distill Llama 70B70B · MMLU-Pro 70.0Q2_K · ~1.1 t/s
- Llama 3.1 70B Instruct70B · MMLU-Pro 66.4Q2_K · ~1.1 t/s
- Mixtral 8x7B Instruct v0.146.7B · MMLU-Pro 29.7NVFP4 · ~1.5 t/s
- Command-R 35B35B · MMLU-Pro 33.0NVFP4 · ~1.2 t/s
- Qwen 3.5 35B-A3B (MoE)35B · MMLU-Pro 84.2NVFP4 · ~6.7 t/s
- Qwen 3.6 35B35B · MMLU-Pro 85.2NVFP4 · ~1.9 t/s
- Yi 1.5 34B Chat34.4B · MMLU-Pro 37.0NVFP4 · ~2 t/s
- Qwen3 32B32.8B · MMLU-Pro 65.5NVFP4 · ~2.2 t/s
- Qwen 2.5 32B Instruct32.5B · MMLU-Pro 69.0NVFP4 · ~2.1 t/s
- Qwen 2.5 Coder 32B Instruct32.5B · MMLU-Pro 50.4NVFP4 · ~2.1 t/s
- DeepSeek R1 Distill Qwen 32B32.5B · MMLU-Pro 65.0NVFP4 · ~2.1 t/s
- Nemotron 3 Nano 30B32B · MMLU-Pro 78.3NVFP4 · ~7.6 t/s
- Gemma 4 31B31B · MMLU-Pro 85.2NVFP4 · ~2 t/s
- Qwen3 30B-A3B (MoE)30B · MMLU-Pro 61.5NVFP4 · ~7.3 t/s
- Gemma 2 27B Instruct27.2B · MMLU-Pro 38.0NVFP4 · ~2.5 t/s
- Gemma 3 27B Instruct27B · MMLU-Pro 67.5NVFP4 · ~2.9 t/s
- Qwen 3.6 27B27B · MMLU-Pro 86.2NVFP4 · ~2.9 t/s
- Gemma 4 26B (MoE)26B · MMLU-Pro 82.6NVFP4 · ~5.9 t/s
- Mistral Small 3.1 24B Instruct24B · MMLU-Pro 66.8NVFP4 · ~3.7 t/s
- Mistral Small 22B22.2B · MMLU-Pro 49.2NVFP4 · ~3.9 t/s
- GPT-OSS 20B21B · MMLU-Pro 67.9NVFP4 · ~8.8 t/s
- Qwen3 14B14.8B · MMLU-Pro 61.0NVFP4 · ~11.3 t/s
- Qwen 2.5 14B Instruct14.7B · MMLU-Pro 63.7NVFP4 · ~10.3 t/s
- Gemma 4 12B (Unified)12B · MMLU-Pro 77.2BF16 · ~1.2 t/s
Too large for this GPU (32)
- Qwen 2.5 72B Instruct
- Mixtral 8x22B Instruct v0.1
- Llama 3.1 405B Instruct
- DeepSeek V3 671B
- DeepSeek R1 671B
- Llama 4 Scout 109B
- Llama 4 Maverick 400B
- Qwen3 235B-A22B (MoE)
- MiniMax M1 456B
- GPT-OSS 120B
- GLM-4.5 355B
- GLM-4.5 Air 106B
- GLM-4.6 355B
- GLM-4.6V 106B
- GLM-4.7 358B
- Qwen 3.5 122B-A10B (MoE)
- MiniMax M2.5 229B
- GLM-5 744B
- MiniMax M2.7 229B
- Nemotron 3 Super 120B
- Kimi K2.6
- GLM-5.1 754B
- DeepSeek V4 Pro 1.6T
- DeepSeek V4 Flash 284B
- GLM-5.2 753B
- Nemotron 3 Ultra 550B-A55B
- Step 3.5 Flash
- Step 3.7 Flash
- MiMo V2.5 Pro
- Kimi K2.5
- MiniMax M3
- Inkling
Continue reading
vram-guides8 min
Best LLMs You Can Run on 8 GB VRAM (2026)tutorials10 min
GGUF Quantization Explained: Q4, Q5, Q6, Q8 Comparedtutorials9 min
Getting Started with Ollama: Run Any LLM in One Commandcomparisons9 min
GGUF vs EXL2 vs AWQ: Which Quantization Format to Usevram-guides8 min
Best LLMs for 6 GB VRAM (2026)tutorials9 min
Getting Started with LM Studio: GUI-Based Local LLMscomparisons10 min
Ollama vs LM Studio vs llama.cpp: Which Tool Should You Use?hardware9 min
Best GPUs Under $500 for LLMs (2026)hardware10 min
NVIDIA RTX 40 Series for LLMs: Complete VRAM & Performance Guidetutorials11 min
Getting Started with Local LLMs (2026): The Complete Beginner's GuideFrequently asked questions
- How much VRAM does the NVIDIA RTX 4060 have?
- The NVIDIA RTX 4060 has 8 GB of GDDR6 with 272 GB/s memory bandwidth.
- What is the NVIDIA RTX 4060 best for?
- With 8 GB of VRAM, the NVIDIA RTX 4060 is best for running compact models (1B–8B) at low quantization, suitable for edge inference, prototyping, and lightweight tasks.
- What LLMs can the NVIDIA RTX 4060 run locally?
- The NVIDIA RTX 4060 can run 23 of the 80 open-weight models tracked by CanItRun natively in VRAM at 8k context. Top options include: Llama 3.1 8B Instruct at NVFP4, Llama 3.2 3B Instruct at NVFP4, Llama 3.2 1B Instruct at FP32.
- Can the NVIDIA RTX 4060 run Llama 3.3 70B Instruct?
- The NVIDIA RTX 4060 can run Llama 3.3 70B Instruct with CPU offload at Q2_K quantization, but inference will be slower than native VRAM execution.
- Can the NVIDIA RTX 4060 run Qwen 3.6 27B?
- The NVIDIA RTX 4060 can run Qwen 3.6 27B with CPU offload at NVFP4 quantization, but inference will be slower than native VRAM execution.
- Can the NVIDIA RTX 4060 run Llama 3.1 8B Instruct?
- Yes. The NVIDIA RTX 4060 runs Llama 3.1 8B Instruct natively in VRAM at NVFP4 quantization, achieving approximately 34.8 tokens per second.