Will this AI model run on my GPU?
VRAM needs for 285 open models on 65 GPUs and Macs, using real GGUF file sizes and each model's own config. Plus live prices for 462 API models, with every price change logged. New models appear within hours of release.
Popular GPUs and Macs
RTX 3060 12GB
12 GB · 171 models fit at 8K contextRTX 4060 Ti 16GB
16 GB · 194 models fit at 8K contextRTX 3090
24 GB · 202 models fit at 8K contextRTX 4090
24 GB · 202 models fit at 8K contextRTX 5090
32 GB · 207 models fit at 8K contextRTX 5070 Ti
16 GB · 194 models fit at 8K contextRX 7900 XTX
24 GB · 202 models fit at 8K contextMac M4 Pro 48GB
48 GB · 207 models fit at 8K contextMac M4 Max 128GB
128 GB · 227 models fit at 8K contextMac M3 Ultra 512GB
512 GB · 278 models fit at 8K contextNewest open models
| Model | Params | Q4_K_M · 8K | On HF since |
|---|---|---|---|
| SingVL S1 4B inclusionAI | 4.5B | 3.6 GB | Nov 11, 2026 |
| Nex N2.6 nex-agi | 321B | 188 GB | Nov 10, 2026 |
| MiniMax H3 Image PE MiniMaxAI | 27B | 16.8 GB | Nov 8, 2026 |
| Qwen3.8 27B DSpine b16 tencent | 2.0B | 2.2 GB | Nov 7, 2026 |
| Arete 27B NousResearch | 27B | 17.4 GB | Nov 6, 2026 |
| Inkling 2 Small thinkingmachines | 594B | 346 GB | Nov 4, 2026 |
| LongCat 2.5 meituan-longcat | 1.78T | 1014 GB | Nov 4, 2026 |
| S5 nano maritaca-ai | 4.7B | 3.7 GB | Nov 2, 2026 |
| Kimi K3.1 Pro moonshotai | 2.78T | 1584 GB | Nov 2, 2026 |
| Mage VL Cowork microsoft | 4.7B | 4.3 GB | Oct 30, 2026 |
Newest API models
| Model | Input / 1M | Output / 1M | Listed |
|---|---|---|---|
| Tencent: Hy4 flash | $0.035 | $0.141 | Nov 16, 2026 |
| Qwen: Qwen4 Max | $1.425 | $4.275 | Nov 16, 2026 |
| inclusionAI: Ling 3.2 Flash (free) | free | free | Nov 13, 2026 |
| Google: Gemini 4 Hydrogen | $1.50 | $7.50 | Nov 12, 2026 |
| Google: Gemini 4 Hydrogen (batch) | $0.75 | $3.75 | Nov 12, 2026 |
| Baidu: ERNIE-5.5-Flash-Preview (free) | free | free | Nov 10, 2026 |
| Baidu: ERNIE-Light-DT-1.0 | $1.13 | $3.94 | Nov 10, 2026 |
| Google: Gemini 4 Carbon | $2.00 | $10.00 | Nov 10, 2026 |
All 462 API prices → · Price change log
Why trust these numbers
Every figure is computed, not guessed: weights from real GGUF file sizes, KV cache from each model's config.json (handles GQA, MLA, sliding-window and hybrid attention, and MoE active vs. total parameters), plus a published overhead. The formulas are on the methodology page, and automated checks hold back any page whose numbers fail a sanity test.