Planner
Build an AI PC part by part
Pick GPU, CPU, RAM, PSU and case. Get live VRAM fit, tok/s, PSU and clearance checks, and a share link.
Open the plannerReverse lookupWhat can my rig run?
Pick the GPU or Mac you already own (and optional multi-GPU count) to see every model that fits, with tok/s.
Check your rigRecipesReady-made builds
Priced part lists for common goals, from a 24 GB budget box to dual 3090s for 70B models.
See buildsPopular checks
Popular models
- Qwen3.8-27B~20.4 GB at Q4
- Qwen3-14B~11.3 GB at Q4
- gpt-oss-20b~15.6 GB at Q4
- Gemma 4 31B~22.8 GB at Q4
- Llama-3.3-70B-Instruct~50.2 GB at Q4
- Qwen3-30B-A3B~22.3 GB at Q4
By VRAM size
- Best LLMs for 6 GB VRAM198 fit
- Best LLMs for 8 GB VRAM315 fit
- Best LLMs for 12 GB VRAM350 fit
- Best LLMs for 16 GB VRAM388 fit
- Best LLMs for 24 GB VRAM423 fit
- Best LLMs for 32 GB VRAM504 fit
- All memory sizes4–128 GB
- VRAM calculatorany size
- How much VRAM do I need?1B–1T
Buying guides
- Best GPU for Llama3 models
- Best GPU for Qwen34 models
- Best GPU for gpt-oss2 models
- Best GPU for DeepSeek R14 models
- Best GPU for Gemma 33 models
- Best GPU for Gemma 45 models
- Best GPU for Qwen3.83 models
- Best GPU for Ornith 1.53 models
- Best GPU for Nemotron3 models
- Best GPU for Liquid AI LFM2 models
- Best GPU for Mistral4 models
- Best GPU for GLM4 models
- Best GPU for Phi4 models
- Best GPU for long-memory models3 models
- Best GPU for local coding models8 models
- Best GPU for local LLMsevery card ranked
- Best Mac for local LLMsMac mini to Mac Studio
- Best mini PC for local LLMsMac mini, Framework, DGX Spark
- Best used GPU for local LLMsRTX 3090, P40, MI50, A6000
- Best AMD GPU for local LLMsRadeon and ROCm support
- Best Intel GPU for local LLMsArc B580, Pro B70, A770
- NVIDIA vs AMD vs Macwhich platform to buy
- Tokens per secondhow fast it runs
- Multi-GPU for local LLMsdual GPU vs one card
- Q4 vs Q8 vs FP16which quant to run
- Context length and VRAM8k to 128k memory
- Ollama VRAM requirementsmemory per Ollama tag
- GGUF VRAM requirementsQ3_K_M to Q8_0 sizes
- MoE vs dense modelsmemory vs speed
- GPU power consumptionwatts and cost per token
- Best GPU by budget$300–$3,500
- GPU vs GPU comparisonshead-to-head
Just added
- LongCat-2.5~1241.8 GB at Q4
- NorthCUA-Pro~277.6 GB at Q4
- ERNIE-5.5-Flash-Base~200.6 GB at Q4
- Apex-35B-V1.0~26.1 GB at Q4
- NorthCUA-Mini~25.5 GB at Q4
- Clover-35B-A3B-Preview~25.2 GB at Q4
- Arete-27B~20.1 GB at Q4
- Groot-19B-A3B-261023~14.5 GB at Q4