NVIDIA RTX 5060
The NVIDIA RTX 5060 has 8 GB VRAM and 448 GB/s memory bandwidth. It can run 25 of our 84 tracked models natively in VRAM at 8k context.
With 8 GB GDDR7, the NVIDIA RTX 5060 is a consumer-tier GPU that can run 25 models natively. It's best for smaller models under 8B parameters.
The NVIDIA RTX 5060 is the entry-level Blackwell GPU with 8GB GDDR7 on a 128-bit bus (448 GB/s) and 3,840 CUDA cores. It is strictly a 1080p gaming card; for LLM inference, only small models like Gemma 4 E4B or Phi-3 Mini fit comfortably in VRAM.
NVIDIA RTX 5060: May 2025 Blackwell GB206 die with 8GB GDDR7 on a 128-bit bus at 448 GB/s — $299 MSRP, entry-level Blackwell.
7B models fit at Q4 with tight context headroom. Larger models require CPU offload. ~7-11 t/s for 7B Q4.
Full CUDA support. 8GB is workable for small models only — treat it as a gaming-first card that can run LLMs, not the reverse.
| Vendor | NVIDIA |
| Architecture | Blackwell |
| VRAM | 8 GB |
| Memory type | GDDR7 |
| Memory bandwidth | 448 GB/s |
| Compute backend | CUDA |
| Tier | Consumer |
| Released | 2025 |
| Models (native) | 25 / 84 |
| Models (offload) | 25 / 84 |
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 (25)
- Bonsai 27B27B · MMLU-Pro 81.51-bit (Q1_0) · ~53.8 t/s
- Phi-4 14B Instruct14B · MMLU-Pro 70.4Q2_K · ~43.6 t/s
- Mistral Nemo 12B Instruct12.2B · MMLU-Pro 35.6Q2_K · ~48.6 t/s
- Gemma 3 12B Instruct12.2B · MMLU-Pro 60.6Q2_K · ~51 t/s
- Gemma 2 9B Instruct9.2B · MMLU-Pro 32.0Q2_K · ~46 t/s
Show 20 more
- Qwen 3.5 9B9B · MMLU-Pro 82.5NVFP4 · ~61.1 t/s
- Llama 3.1 8B Instruct8B · MMLU-Pro 48.3NVFP4 · ~57.4 t/s
- DeepSeek R1 Distill Llama 8B8B · MMLU-Pro 41.0NVFP4 · ~57.4 t/s
- Qwen3 8B8B · MMLU-Pro 56.7NVFP4 · ~55.9 t/s
- Qwen 2.5 7B Instruct7.6B · MMLU-Pro 56.3NVFP4 · ~68.2 t/s
- Mistral 7B Instruct v0.37.25B · MMLU-Pro 30.0NVFP4 · ~62 t/s
- Gemma 3 4B Instruct4B · MMLU-Pro 43.6NVFP4 · ~116.3 t/s
- Gemma 4 E4B4B · MMLU-Pro 69.4NVFP4 · ~96.9 t/s
- Phi-3.5 Mini Instruct3.8B · MMLU-Pro 47.4NVFP4 · ~56.9 t/s
- Phi-4-mini Instruct3.8B · MMLU-Pro 67.3NVFP4 · ~97.9 t/s
- Llama 3.2 3B Instruct3.2B · MMLU-Pro 24.0NVFP4 · ~114.7 t/s
- Qwen 2.5 3B Instruct3.1B · MMLU-Pro 32.4BF16 · ~44.8 t/s
- Gemma 2 2B Instruct2.6B · MMLU-Pro 17.8BF16 · ~48 t/s
- Gemma 4 E2B2B · MMLU-Pro 60.0BF16 · ~66.1 t/s
- SmolLM2 1.7B Instruct1.7B · MMLU-Pro 19.0BF16 · ~58.1 t/s
- Qwen 2.5 1.5B Instruct1.5B · MMLU-Pro 16.8FP32 · ~46.7 t/s
- Llama 3.2 1B Instruct1.24B · MMLU-Pro 12.5FP32 · ~55.7 t/s
- Gemma 3 1B Instruct1B · MMLU-Pro 14.7FP32 · ~67.3 t/s
- Qwen 2.5 0.5B Instruct0.5B · MMLU-Pro 10.0FP32 · ~138.6 t/s
- SmolLM2 360M Instruct0.36B · MMLU-Pro 8.0FP32 · ~164 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 85.3NVFP4 · ~8 t/s
- Qwen 3.6 35B35B · MMLU-Pro 85.2NVFP4 · ~2 t/s
- Yi 1.5 34B Chat34.4B · MMLU-Pro 37.0NVFP4 · ~2 t/s
- Qwen3 32B32.8B · MMLU-Pro 65.5NVFP4 · ~2.3 t/s
- Qwen 2.5 32B Instruct32.5B · MMLU-Pro 69.0NVFP4 · ~2.2 t/s
- Qwen 2.5 Coder 32B Instruct32.5B · MMLU-Pro 50.4NVFP4 · ~2.2 t/s
- DeepSeek R1 Distill Qwen 32B32.5B · MMLU-Pro 65.0NVFP4 · ~2.2 t/s
- Nemotron 3 Nano 30B32B · MMLU-Pro 78.3NVFP4 · ~8 t/s
- Gemma 4 31B31B · MMLU-Pro 85.2NVFP4 · ~2.1 t/s
- Qwen3 30B-A3B (MoE)30B · MMLU-Pro 61.5NVFP4 · ~7.7 t/s
- Gemma 2 27B Instruct27.2B · MMLU-Pro 38.0NVFP4 · ~2.6 t/s
- Gemma 3 27B Instruct27B · MMLU-Pro 67.5NVFP4 · ~3.1 t/s
- Qwen 3.6 27B27B · MMLU-Pro 86.2NVFP4 · ~3.5 t/s
- Gemma 4 26B (MoE)26B · MMLU-Pro 82.6NVFP4 · ~7.8 t/s
- Mistral Small 3.1 24B Instruct24B · MMLU-Pro 66.8NVFP4 · ~3.9 t/s
- Mistral Small 22B22.2B · MMLU-Pro 49.2NVFP4 · ~4.1 t/s
- GPT-OSS 20B21B · MMLU-Pro 67.9NVFP4 · ~9.9 t/s
- Qwen3 14B14.8B · MMLU-Pro 61.0NVFP4 · ~13.9 t/s
- Qwen 2.5 14B Instruct14.7B · MMLU-Pro 63.7NVFP4 · ~12.4 t/s
- Gemma 4 12B (Unified)12B · MMLU-Pro 77.2BF16 · ~1.2 t/s
Too large for this GPU (34)
- 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
- Mistral Medium 3.5 128B
- 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
- Kimi K3
Continue reading
Frequently asked questions
- How much VRAM does the NVIDIA RTX 5060 have?
- The NVIDIA RTX 5060 has 8 GB of GDDR7 with 448 GB/s memory bandwidth.
- What is the NVIDIA RTX 5060 best for?
- With 8 GB of VRAM, the NVIDIA RTX 5060 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 5060 run locally?
- The NVIDIA RTX 5060 can run 25 of the 84 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 5060 run Llama 3.3 70B Instruct?
- The NVIDIA RTX 5060 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 5060 run Qwen 3.6 27B?
- The NVIDIA RTX 5060 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 5060 run Llama 3.1 8B Instruct?
- Yes. The NVIDIA RTX 5060 runs Llama 3.1 8B Instruct natively in VRAM at NVFP4 quantization, achieving approximately 57.4 tokens per second.