# How much VRAM do you need for a 16B–21B model? > A 16B–21B LLM needs about 15.6 GB of VRAM at Q4, 26.6 GB at Q8 and 49.1 GB at FP16. Context-length table and the cheapest GPU and Mac that fit. Source: https://whatcanirunai.vercel.app/how-much-vram/20b/ # How much VRAM do you need for a 16B–21B model? Short answer: about **15.6 GB** at Q4\_K\_M, the usual quant for local use. Below: every quant, longer contexts, the cheapest hardware that fits, and every 16B–21B open model we track. Q4\_K\_M 15.6 GB 12.7 GB weights Q8\_0 26.6 GB 22.2 GB weights FP16 49.1 GB 41.8 GB weights Worked out for [gpt-oss-20b](https://whatcanirunai.vercel.app/models/gpt-oss-20b/) (20.9B parameters, from the Hugging Face API). Formula: weights × 1.15 + 1 GB at a \~4k context ([how we calculate](https://whatcanirunai.vercel.app/method/)). Cheapest GPU that fits at Q4: **Radeon RX 7900 XT** (20 GB, $899). Cheapest Mac or mini PC: **Mac mini M6 24GB** ($1,299). ## Cheapest hardware for each quant Only devices with a sourced price. GPUs at list price (MSRP); street prices may be higher. __Cheapest GPU and Mac or mini PC for a 16B–21B model at each quant__ | Setting | Memory needed | Cheapest GPU | Cheapest Mac / mini PC | | ---------------------- | ------------- | --------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------- | | Q4\_K\_M, \~4k context | 15.6 GB | [Radeon RX 7900 XT](https://whatcanirunai.vercel.app/hardware/rx-7900-xt/) 20 GB usable · $899 | [Mac mini M6 24GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-24/) 18 GB usable · $1,299 | | Q8\_0, \~4k context | 26.6 GB | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/) 32 GB usable · $949 | [Mac mini M5 Pro 64GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m5-pro-64/) 48 GB usable · $3,199 | | FP16, \~4k context | 49.1 GB | [NVIDIA RTX Pro 6000 Blackwell 96GB](https://whatcanirunai.vercel.app/hardware/rtx-pro-6000-blackwell/) 96 GB usable · $8,565 | [Framework Desktop (Ryzen AI Max+ 395) 128GB](https://whatcanirunai.vercel.app/hardware/framework-desktop-128/) 96 GB usable · $3,449 | | Q4\_K\_M, 32k context | 16.3 GB | [Radeon RX 7900 XT](https://whatcanirunai.vercel.app/hardware/rx-7900-xt/) 20 GB usable · $899 | [Mac mini M6 32GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-32/) 24 GB usable · $1,799 | ## Memory by context length Each extra token of context adds to the KV cache. Numbers for gpt-oss-20b, from its published config; a Q8 KV cache halves the cache part. __Memory needed for a 16B–21B model by context length__ | Context | Q4\_K\_M, FP16 cache | Q4\_K\_M, Q8 cache | Q8\_0, FP16 cache | | ----------- | -------------------- | ------------------ | ----------------- | | 4k tokens | 15.6 GB | 15.6 GB | 26.6 GB | | 8k tokens | 15.7 GB | 15.6 GB | 26.7 GB | | 16k tokens | 15.9 GB | 15.7 GB | 26.9 GB | | 32k tokens | 16.3 GB | 15.9 GB | 27.3 GB | | 64k tokens | 17.1 GB | 16.3 GB | 28.1 GB | | 128k tokens | 18.7 GB | 17.1 GB | 29.7 GB | ## Hardware that runs it Bar: Q4 memory needed vs. usable memory. Speed is estimated decode at Q4 ([how we estimate](https://whatcanirunai.vercel.app/method/#speed)); MoE shows an upper bound only.FitsTight __Tracked hardware that runs a 16B–21B model at Q4__ | Hardware | Q4 need vs. usable | Q4 / Q8 / FP16 | Est. tok/s | | ------------------------------------------------------------------------------------------------------ | ------------------- | ------------------------------- | ---------- | | [Radeon RX 7600 XT](https://whatcanirunai.vercel.app/hardware/rx-7600-xt/)GPU · $329 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 112 | | [Radeon RX 9060 XT 16GB](https://whatcanirunai.vercel.app/hardware/rx-9060-xt-16gb/)GPU · $349 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 125 | | [Intel Arc A770 16GB](https://whatcanirunai.vercel.app/hardware/arc-a770-16gb/)GPU · $349 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 218 | | [GeForce RTX 5060 Ti 16GB](https://whatcanirunai.vercel.app/hardware/rtx-5060-ti-16gb/)GPU · $429 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 174 | | [GeForce RTX 4060 Ti 16GB](https://whatcanirunai.vercel.app/hardware/rtx-4060-ti-16gb/)GPU · $499 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 112 | | [Radeon RX 7800 XT](https://whatcanirunai.vercel.app/hardware/rx-7800-xt/)GPU · $499 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 243 | | [Radeon RX 7900 GRE](https://whatcanirunai.vercel.app/hardware/rx-7900-gre/)GPU · $549 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 224 | | [Radeon RX 9070](https://whatcanirunai.vercel.app/hardware/rx-9070/)GPU · $549 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 249 | | [Radeon RX 6800](https://whatcanirunai.vercel.app/hardware/rx-6800/)GPU · $579 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 199 | | [Radeon RX 9070 XT](https://whatcanirunai.vercel.app/hardware/rx-9070-xt/)GPU · $599 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 249 | | [Radeon RX 6800 XT](https://whatcanirunai.vercel.app/hardware/rx-6800-xt/)GPU · $649 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 199 | | [NVIDIA RTX 2000 Ada 16GB](https://whatcanirunai.vercel.app/hardware/rtx-2000-ada/)GPU · $649 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 87 | | [GeForce RTX 5070 Ti](https://whatcanirunai.vercel.app/hardware/rtx-5070-ti/)GPU · $749 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 349 | | [GeForce RTX 4070 Ti Super](https://whatcanirunai.vercel.app/hardware/rtx-4070-ti-super/)GPU · $799 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 262 | | [GeForce RTX 4080 Super](https://whatcanirunai.vercel.app/hardware/rtx-4080-super/)GPU · $999 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 287 | | [GeForce RTX 5080](https://whatcanirunai.vercel.app/hardware/rtx-5080/)GPU · $999 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 374 | | [Radeon RX 6900 XT](https://whatcanirunai.vercel.app/hardware/rx-6900-xt/)GPU · $999 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 199 | | [NVIDIA RTX A4000](https://whatcanirunai.vercel.app/hardware/rtx-a4000/)GPU · $1,000 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 174 | | [Radeon RX 6950 XT](https://whatcanirunai.vercel.app/hardware/rx-6950-xt/)GPU · $1,099 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 224 | | [GeForce RTX 4080](https://whatcanirunai.vercel.app/hardware/rtx-4080/)GPU · $1,199 | **15.6** / 16.0 GB | Q4TightQ8Won’t fitFP16Won’t fit | up to 279 | | [Mac mini M6 24GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-24/)unified memory · $1,299 | **15.6** / 18.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 62 | | [Radeon RX 7900 XT](https://whatcanirunai.vercel.app/hardware/rx-7900-xt/)GPU · $899 | **15.6** / 20.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 312 | | [NVIDIA RTX 4000 Ada 20GB](https://whatcanirunai.vercel.app/hardware/rtx-4000-ada/)GPU · $1,250 | **15.6** / 20.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 140 | | [NVIDIA Tesla P40 24GB](https://whatcanirunai.vercel.app/hardware/tesla-p40/)GPU | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 135 | | [Radeon RX 7900 XTX](https://whatcanirunai.vercel.app/hardware/rx-7900-xtx/)GPU · $999 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 374 | | [GeForce RTX 3090](https://whatcanirunai.vercel.app/hardware/rtx-3090/)GPU · $1,499 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 365 | | [GeForce RTX 4090](https://whatcanirunai.vercel.app/hardware/rtx-4090/)GPU · $1,599 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 393 | | [Mac mini M6 32GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m6-32/)unified memory · $1,799 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 62 | | [GeForce RTX 3090 Ti](https://whatcanirunai.vercel.app/hardware/rtx-3090-ti/)GPU · $1,999 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 393 | | [NVIDIA RTX A5000](https://whatcanirunai.vercel.app/hardware/rtx-a5000/)GPU · $2,250 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 299 | | [MacBook Pro M6 32GB](https://whatcanirunai.vercel.app/hardware/macbook-pro-m6-32/)unified memory · $2,399 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 62 | | [NVIDIA Titan RTX](https://whatcanirunai.vercel.app/hardware/titan-rtx/)GPU · $2,499 | **15.6** / 24.0 GB | Q4FitsQ8Won’t fitFP16Won’t fit | up to 262 | | [NVIDIA Tesla V100 32GB (PCIe)](https://whatcanirunai.vercel.app/hardware/tesla-v100-32gb/)GPU | **15.6** / 32.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 351 | | [AMD Radeon Instinct MI50 32GB](https://whatcanirunai.vercel.app/hardware/instinct-mi50-32gb/)GPU | **15.6** / 32.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 399 | | [AMD Radeon Instinct MI60 32GB](https://whatcanirunai.vercel.app/hardware/instinct-mi60/)GPU | **15.6** / 32.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 399 | | [Intel Arc Pro B70 32GB](https://whatcanirunai.vercel.app/hardware/arc-pro-b70/)GPU · $949 | **15.6** / 32.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 237 | | [GeForce RTX 5090](https://whatcanirunai.vercel.app/hardware/rtx-5090/)GPU · $1,999 | **15.6** / 32.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 698 | | [Mac mini M5 Pro 64GB](https://whatcanirunai.vercel.app/hardware/mac-mini-m5-pro-64/)unified memory · $3,199 | **15.6** / 48.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 113 | | [NVIDIA RTX A6000](https://whatcanirunai.vercel.app/hardware/rtx-a6000/)GPU · $4,649 | **15.6** / 48.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 299 | | [NVIDIA DGX Spark 64GB](https://whatcanirunai.vercel.app/hardware/dgx-spark-64/)unified memory · $4,999 | **15.6** / 48.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 100 | | [NVIDIA RTX 6000 Ada 48GB](https://whatcanirunai.vercel.app/hardware/rtx-6000-ada/)GPU · $6,799 | **15.6** / 48.0 GB | Q4FitsQ8FitsFP16Won’t fit | up to 374 | | [Framework Desktop (Ryzen AI Max+ 395) 128GB](https://whatcanirunai.vercel.app/hardware/framework-desktop-128/)unified memory · $3,449 | **15.6** / 96.0 GB | Q4FitsQ8FitsFP16Fits | up to 94 | | [MacBook Pro M4 Max 128GB](https://whatcanirunai.vercel.app/hardware/macbook-pro-m4-max-128/)unified memory · $4,999 | **15.6** / 96.0 GB | Q4FitsQ8FitsFP16Fits | up to 200 | | [MacBook Pro M5 Max 128GB](https://whatcanirunai.vercel.app/hardware/macbook-pro-m5-max-128/)unified memory · $6,699 | **15.6** / 96.0 GB | Q4FitsQ8FitsFP16Fits | up to 225 | | [NVIDIA DGX Spark 128GB](https://whatcanirunai.vercel.app/hardware/dgx-spark-128/)unified memory · $6,950 | **15.6** / 96.0 GB | Q4FitsQ8FitsFP16Fits | up to 100 | | [NVIDIA RTX Pro 6000 Blackwell 96GB](https://whatcanirunai.vercel.app/hardware/rtx-pro-6000-blackwell/)GPU · $8,565 | **15.6** / 96.0 GB | Q4FitsQ8FitsFP16Fits | up to 698 | | [Mac Studio M5 Ultra 256GB](https://whatcanirunai.vercel.app/hardware/mac-studio-m5-ultra-256/)unified memory · $9,499 | **15.6** / 192.0 GB | Q4FitsQ8FitsFP16Fits | up to 440 | | [Mac Studio M5 Ultra 512GB](https://whatcanirunai.vercel.app/hardware/mac-studio-m5-ultra-512/)unified memory · $22,299 | **15.6** / 384.0 GB | Q4FitsQ8FitsFP16Fits | up to 440 | ## 16B–21B models we track __Memory needed by each 16B–21B model we track__ | Model | Q4\_K\_M | Q8\_0 | FP16 | | ----------------------------------------------------------------------------------------------- | -------- | ------- | ------- | | [LLaDA2.1-mini](https://whatcanirunai.vercel.app/models/llada2.1-mini/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [LLaDA2.0-mini](https://whatcanirunai.vercel.app/models/llada2.0-mini/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [Ling-mini-2.0](https://whatcanirunai.vercel.app/models/ling-mini-2.0/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [Ring-mini-2.0](https://whatcanirunai.vercel.app/models/ring-mini-2.0/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [LLaDA2.2-mini](https://whatcanirunai.vercel.app/models/llada2.2-mini/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [LLaDA2.0-mini-preview](https://whatcanirunai.vercel.app/models/llada2.0-mini-preview/)16.3B params · inclusionAI · MoE | 12.3 GB | 20.9 GB | 38.4 GB | | [Kimi-VL-A3B-Instruct](https://whatcanirunai.vercel.app/models/kimi-vl-a3b-instruct/)16.4B params · moonshotai · MoE | 12.4 GB | 21.0 GB | 38.7 GB | | [Kimi-VL-A3B-Thinking](https://whatcanirunai.vercel.app/models/kimi-vl-a3b-thinking/)16.4B params · moonshotai · MoE | 12.4 GB | 21.0 GB | 38.7 GB | | [Kimi-VL-A3B-Thinking-2506](https://whatcanirunai.vercel.app/models/kimi-vl-a3b-thinking-2506/)16.4B params · moonshotai · MoE | 12.4 GB | 21.0 GB | 38.7 GB | | [Ring-mini-linear-2.0](https://whatcanirunai.vercel.app/models/ring-mini-linear-2.0/)16.4B params · inclusionAI · MoE | 12.5 GB | 21.1 GB | 38.8 GB | | [Ling-lite](https://whatcanirunai.vercel.app/models/ling-lite/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Ling-lite-1.5-2507](https://whatcanirunai.vercel.app/models/ling-lite-1.5-2507/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Ling-lite-1.5](https://whatcanirunai.vercel.app/models/ling-lite-1.5/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Ring-lite](https://whatcanirunai.vercel.app/models/ring-lite/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Ling-Coder-lite](https://whatcanirunai.vercel.app/models/ling-coder-lite/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Ring-lite-distill-preview](https://whatcanirunai.vercel.app/models/ring-lite-distill-preview/)16.8B params · inclusionAI · MoE | 12.7 GB | 21.5 GB | 39.6 GB | | [Groot-19B-A3B-261023](https://whatcanirunai.vercel.app/models/groot-19b-a3b-261023/)19.4B params · GrootTeam · MoE | 14.5 GB | 24.7 GB | 45.7 GB | | [internlm2\_5-20b-chat](https://whatcanirunai.vercel.app/models/internlm2-5-20b-chat/)19.9B params · internlm | 14.8 GB | 25.3 GB | 46.7 GB | | [maple-preview](https://whatcanirunai.vercel.app/models/maple-preview/)20.2B params · deepgrove · MoE | 15.1 GB | 25.7 GB | 47.5 GB | | [reka-flash-3](https://whatcanirunai.vercel.app/models/reka-flash-3/)20.9B params · RekaAI | 15.6 GB | 26.5 GB | 49.1 GB | | [gpt-oss-20b](https://whatcanirunai.vercel.app/models/gpt-oss-20b/)20.9B params · openai · MoE | 15.6 GB | 26.6 GB | 49.1 GB | | [ERNIE-4.5-21B-A3B-Thinking](https://whatcanirunai.vercel.app/models/ernie-4.5-21b-a3b-thinking/)21.8B params · baidu · MoE | 16.2 GB | 27.7 GB | 51.2 GB | | [ERNIE-4.5-21B-A3B-PT](https://whatcanirunai.vercel.app/models/ernie-4.5-21b-a3b-pt/)21.9B params · baidu · MoE | 16.3 GB | 27.8 GB | 51.5 GB | Have a different amount of memory? Try the [VRAM calculator](https://whatcanirunai.vercel.app/vram-calculator/) or see [best LLMs by VRAM size](https://whatcanirunai.vercel.app/best-llm-for/). ## Quick answers How much VRAM do I need for a 16B–21B model? About 15.6 GB at Q4\_K\_M, 26.6 GB at Q8\_0 and 49.1 GB at FP16, at a \~4k context. That is for gpt-oss-20b (20.9B parameters): weights plus about 15% for the KV cache and 1 GB for the runtime. How much VRAM does a 16B–21B model need for 32k context? About 16.3 GB at Q4\_K\_M with an FP16 KV cache, or 15.9 GB with a Q8 cache, based on gpt-oss-20b's published config. Models with different attention layouts need more or less. What is the cheapest GPU for a 16B–21B model? Of the GPUs we track, the Radeon RX 7900 XT (20 GB, $899 list price) is the cheapest that runs it fully on the GPU at Q4. Can a Mac run a 16B–21B model? Yes. The cheapest Mac or unified-memory mini PC we track that runs it at Q4 is the Mac mini M6 24GB (24 GB, about 18.0 GB usable, $1,299). Does a MoE model of the same size need less VRAM? No. A mixture-of-experts model still has to hold all of its weights in memory, so it needs about the same VRAM as a dense model with the same total parameters. It runs faster, because only a few experts are read per token. Other sizes: [1B–2B](https://whatcanirunai.vercel.app/how-much-vram/1b/) · [2B–3B](https://whatcanirunai.vercel.app/how-much-vram/3b/) · [4B–6B](https://whatcanirunai.vercel.app/how-much-vram/4b/) · [7B–8B](https://whatcanirunai.vercel.app/how-much-vram/8b/) · [12B–14B](https://whatcanirunai.vercel.app/how-much-vram/14b/) · [24B–27B](https://whatcanirunai.vercel.app/how-much-vram/27b/) · [30B–32B](https://whatcanirunai.vercel.app/how-much-vram/32b/) · [40B–50B](https://whatcanirunai.vercel.app/how-much-vram/49b/) · [70B](https://whatcanirunai.vercel.app/how-much-vram/70b/) · [120B](https://whatcanirunai.vercel.app/how-much-vram/120b/) · [200B–300B](https://whatcanirunai.vercel.app/how-much-vram/235b/) · [300B–480B](https://whatcanirunai.vercel.app/how-much-vram/400b/) · [500B–800B](https://whatcanirunai.vercel.app/how-much-vram/671b/) · [1T+](https://whatcanirunai.vercel.app/how-much-vram/1t/) · [All sizes](https://whatcanirunai.vercel.app/how-much-vram/) · [Ollama VRAM by tag](https://whatcanirunai.vercel.app/ollama-vram-requirements/) · [GGUF sizes by quant](https://whatcanirunai.vercel.app/gguf-vram-requirements/) --- Agent guide: https://whatcanirunai.vercel.app/llms.txt · JSON API: https://whatcanirunai.vercel.app/developers/