Can I run it?
48 open models × 128 machines. Every answer comes from the same engine as the full check: memory, speed and context for real workloads, not just “does the file fit”.
Popular machines
How many catalog models each one runs acceptably or better for chat.
- MacBook Air M4 16 GB16 GB unified11 of 48 models run well
- MacBook Air M5 24 GB24 GB unified25 of 48 models run well
- MacBook Pro M4 Pro 24 GB24 GB unified25 of 48 models run well
- MacBook Pro M4 Pro 48 GB48 GB unified33 of 48 models run well
- MacBook Pro M4 Max 128 GB128 GB unified41 of 48 models run well
- MacBook Pro M5 Pro 48 GB48 GB unified33 of 48 models run well
- Mac mini M4 16 GB16 GB unified11 of 48 models run well
- Mac mini M4 Pro 64 GB64 GB unified37 of 48 models run well
- Mac Studio M4 Max 128 GB128 GB unified41 of 48 models run well
- Mac Studio M3 Ultra 512 GB512 GB unified47 of 48 models run well
- Desktop PC · GeForce RTX 3060 12GB · 32 GB RAM12 GB VRAM + 32 GB RAM13 of 48 models run well
- Desktop PC · GeForce RTX 3090 · 64 GB RAM24 GB VRAM + 64 GB RAM31 of 48 models run well
- Desktop PC · GeForce RTX 4060 Ti 16GB · 32 GB RAM16 GB VRAM + 32 GB RAM17 of 48 models run well
- Desktop PC · GeForce RTX 4090 · 64 GB RAM24 GB VRAM + 64 GB RAM31 of 48 models run well
- Desktop PC · GeForce RTX 5090 · 64 GB RAM32 GB VRAM + 64 GB RAM31 of 48 models run well
- Desktop PC · Radeon RX 7900 XTX · 64 GB RAM24 GB VRAM + 64 GB RAM31 of 48 models run well
- Laptop · GeForce RTX 4060 Laptop · 32 GB RAM8 GB VRAM + 32 GB RAM10 of 48 models run well
- Ryzen AI Max+ 395 mini PC / Framework Desktop · 128 GB128 GB unified40 of 48 models run well
- NVIDIA DGX Spark (GB10) · 128 GB128 GB unified41 of 48 models run well
Models
Every open model we rate, newest first. Each page lists every machine with its verdict.
- DeepSeek-V4.1-Flash (748B-A16B)DeepSeek · 748B (16B active) · 1M contextOpen source
- Kolibri-1 (78B-A3.5B)Aleph Alpha · 78B (3.5B active) · 256K contextOpen source
- LFM2.5 24B-A2BLiquid AI · 24B (2.0B active) · 128K context
- Muse Spark Open 816B-A42BMeta · 816B (42B active) · 1M context
- Nemotron 3.5 Bolt 74B-A9BNVIDIA · 74B (9.0B active) · 256K context
- Nemotron 3.5 Super VL 120B-A12BNVIDIA · 120B (12B active) · 1M context
- Reflection Beam (501B-A23B)Reflection AI · 501B (23B active) · 1M contextOpen source
- Step 5 Preview 600B-A27BStepfun · 600B (27B active) · 1M context
- MiniCPM5 2BOpenBMB · 2.5B · 128K contextOpen source
- GLM-5.3 (753B-A40B)Z.ai · 753B (40B active) · 1M context
- GLM-5.3-Flash (320B-A18B)Z.ai · 320B (18B active) · 1M contextOpen source
- Muse Glimmer (30B)Meta · 30B · 128K contextOpen source
- Qwen3.8 27BAlibaba Qwen · 28B · 256K contextOpen source
- Gemma 4 12BGoogle · 12B · 256K contextOpen source
- DeepSeek-V4-Flash (284B-A13B)DeepSeek · 284B (13B active) · 1M contextOpen source
- Gemma 4 26B-A4BGoogle · 26B (3.8B active) · 256K contextOpen source
- Gemma 4 31BGoogle · 31B · 256K contextOpen source
- Gemma 4 E4BGoogle · 8.0B · 128K contextOpen source
- Qwen3.6 27BAlibaba Qwen · 28B · 256K contextOpen source
- Qwen3.6 35B-A3BAlibaba Qwen · 36B (3.0B active) · 256K contextOpen source
- Mistral Small 4 (119B-A6.5B)Mistral AI · 119B (6.5B active) · 256K contextOpen source
- Qwen3.5 122B-A10BAlibaba Qwen · 125B (10B active) · 256K contextOpen source
- Qwen3.5 27BAlibaba Qwen · 28B · 256K contextOpen source
- Qwen3.5 35B-A3BAlibaba Qwen · 36B (3.0B active) · 256K contextOpen source
- Qwen3.5 4BAlibaba Qwen · 4.7B · 256K contextOpen source
- Qwen3.5 9BAlibaba Qwen · 9.7B · 256K contextOpen source
- GLM-4.7-Flash (30B-A3B)Z.ai · 31B (3.0B active) · 198K contextOpen source
- Devstral Small 2 (24B)Mistral AI · 24B · 256K contextOpen source
- GLM-4.6V-Flash (10B)Z.ai · 10B · 128K contextOpen source
- Qwen3-Next 80B-A3B InstructAlibaba Qwen · 81B (3.0B active) · 256K contextOpen source
- gpt-oss-120bOpenAI · 117B (5.1B active) · 128K contextOpen source
- gpt-oss-20bOpenAI · 21B (3.6B active) · 128K contextOpen source
- GLM-4.5-Air (106B-A12B)Z.ai · 111B (12B active) · 128K contextOpen source
- Qwen3 30B-A3B Instruct 2507Alibaba Qwen · 31B (3.3B active) · 256K contextOpen source
- Qwen3-Coder 30B-A3B InstructAlibaba Qwen · 31B (3.3B active) · 256K contextOpen source
- Mistral Small 3.2 (24B)Mistral AI · 24B · 128K contextOpen source
- DeepSeek-R1-0528-Qwen3-8BDeepSeek · 8.2B · 128K contextOpen source
- Llama 4 Scout (109B-A17B)Meta · 109B (17B active) · 256K contextOpen weights (restricted)
- Qwen3 14BAlibaba Qwen · 15B · 32K contextOpen source
- Qwen3 32BAlibaba Qwen · 33B · 32K contextOpen source
- Qwen3 8BAlibaba Qwen · 8.2B · 32K contextOpen source
- Gemma 3 27BGoogle · 27B · 128K contextOpen weights (restricted)
- Llama 3.3 70B InstructMeta · 71B · 128K contextOpen weights (restricted)
- Phi-4 (14B)Microsoft · 15B · 16K contextOpen source
- Qwen2.5-Coder 32B InstructAlibaba Qwen · 33B · 32K contextOpen source
- Llama 3.2 3B InstructMeta · 3.2B · 128K contextOpen weights (restricted)
- Qwen2.5-Coder 7B InstructAlibaba Qwen · 7.6B · 32K contextOpen source
- Llama 3.1 8B InstructMeta · 8.0B · 128K contextOpen weights (restricted)