Which GPU instance or cluster can actually hold it?
A model fits when its weights, plus the memory its attention cache needs for your context length and the requests you serve at once, are within the GPU memory of an instance or a cluster. Anything too small is left out; what fits is listed per cloud with its price.
Memory needed and the cheapest hardware that fits, by model
| Model | Parameters | Precision | GPU memory needed | Cheapest instance that fits |
|---|---|---|---|---|
| OpenAI gpt-oss 120B | 116.83 B | int4 | 73 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| GLM 5.3 | 753.33 B | fp8 | 941.7 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Kimi K3 | 2779.93 B | published size | 1951.1 GB | OCI BM.GPU.MI355X.8: 8 × MI355X, $68.80 an hour |
| GLM 5.2 | 753.33 B | bf16 | 1883.3 GB | OCI BM.GPU.MI355X.8: 8 × MI355X, $68.80 an hour |
| Llama 3.3 70B | 70.55 B | bf16 | 176.4 GB | Core42 mi300x.1x: 1 × MI300X, $3.50 an hour |
| OpenAI gpt-oss 20B | 20.91 B | int4 | 13.1 GB | HF Endpoints aws.nvidia-t4.x1: 1 × T4, $0.50 an hour |
| GLM 5.3 Flash | 321.32 B | fp8 | 401.6 GB | Crusoe a100-sxm.8x: 8 × A100, $18.40 an hour |
| Llama 4 Maverick 17B | 401.58 B | bf16 | 1003.9 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Qwen3.8 27B | 27.78 B | bf16 | 69.5 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| MiniMax M3 | 427.04 B | bf16 | 1067.6 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| DeepSeek V4 Pro | 1598.84 B | published size | 1080.9 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| DeepSeek V4 Pro 0813 | 1650.5 B | published size | 1115.9 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| DeepSeek V4.1 Flash | 763.21 B | published size | 637.9 GB | Crusoe a100-sxm.8x: 8 × A100, $18.40 an hour |
| Gemma 4 26B A4B | 25.81 B | bf16 | 64.5 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| Llama 3.1 8B | 8.03 B | bf16 | 20.1 GB | HF Endpoints gcp.nvidia-l4.x1: 1 × L4, $0.70 an hour |
| Llama 4 Scout 17B | 108.64 B | bf16 | 271.6 GB | Nebius gpu-b300-sxm.1gpu-24vcpu-346gb: 1 × B300, $9.50 an hour |
| Kimi K2.5 | 1026.88 B | published size | 744 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Kimi K2.6 | 1026.88 B | published size | 744 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Qwen3 32B | 32.76 B | bf16 | 81.9 GB | Nebius gpu-rtx6000-a.1gpu-24vcpu-218gb: 1 × RTX PRO 6000, $1.80 an hour |
| DeepSeek V3.2 | 685.36 B | fp8 | 856.7 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| DeepSeek V4 Flash | 290.94 B | published size | 199.5 GB | DeepInfra b300.1x: 1 × B300, $5.89 an hour |
| DeepSeek V4 Flash 0731 | 304.18 B | published size | 208.6 GB | DeepInfra b300.1x: 1 × B300, $5.89 an hour |
| Llama 3.1 70B | 70.55 B | bf16 | 176.4 GB | Core42 mi300x.1x: 1 × MI300X, $3.50 an hour |
| Mistral Small | 23.57 B | bf16 | 58.9 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| MiniMax M2.5 | 228.69 B | fp8 | 285.9 GB | Nebius gpu-b300-sxm.1gpu-24vcpu-346gb: 1 × B300, $9.50 an hour |
| DeepSeek V3.1 | 684.53 B | fp8 | 855.7 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Gemma 3 27B | 27.43 B | bf16 | 68.6 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| Llama 3.1 405B | 405.85 B | bf16 | 1014.6 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Phi-4 | 14.66 B | bf16 | 36.6 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| Mistral 7B | 7.25 B | bf16 | 18.1 GB | HF Endpoints gcp.nvidia-l4.x1: 1 × L4, $0.70 an hour |
| Kimi K2.7 Code | 1026.88 B | published size | 744 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| GLM 5 | 753.86 B | bf16 | 1884.7 GB | OCI BM.GPU.MI355X.8: 8 × MI355X, $68.80 an hour |
| Hermes 4 405B | 405.85 B | bf16 | 1014.6 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Qwen3 235B A22B Instruct 2507 | 235.09 B | bf16 | 587.7 GB | Crusoe a100-sxm.8x: 8 × A100, $18.40 an hour |
| Qwen3 30B A3B Instruct 2507 | 30.53 B | bf16 | 76.3 GB | DeepInfra a100.1x: 1 × A100, $1.49 an hour |
| Qwen3.5 397B A17B | 403.4 B | bf16 | 1008.5 GB | Crusoe mi300x.8x: 8 × MI300X, $27.60 an hour |
| Mistral Small 4 | 119.4 B | fp8 | 149.2 GB | Core42 mi300x.1x: 1 × MI300X, $3.50 an hour |
| Qwen3 235B A22B | 235.09 B | bf16 | 587.7 GB | Crusoe a100-sxm.8x: 8 × A100, $18.40 an hour |
| GLM 5.1 | 753.86 B | bf16 | 1884.7 GB | OCI BM.GPU.MI355X.8: 8 × MI355X, $68.80 an hour |