What can the M4 Max 128 GB run?
| Text models: 119 | Visual models: 9 | Strongest first |
- Memory
- 128 GB unified
- Bandwidth
- 546 GB/s
- Maker
- Apple
Text models that run on it
| Model | Best fit | Context | Speed | Strongest at |
|---|---|---|---|---|
| Qwen3.8 Flash-Next | 3-bit | 256K | 71 tok/s | Chat, Coding, Agents |
| MiniMax M3 | 1.6-bit | 64K | 50 tok/s | Docs |
| Qwen3.8 27B | 8-bit | 256K | 13 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.5 397B A17B | 1.6-bit | 256K | 57 tok/s | Agents, Docs |
| MiniMax M2.7 | 2-bit | 40K | 56 tok/s | Coding, Agents |
| Muse Glimmer 30B | 8-bit | 128K | 13 tok/s | Docs |
| Mistral Medium 3.5 128B | 5-bit | 16K | 4.4 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.5 122B A10B | 5-bit | 256K | 41 tok/s | Chat, Coding, Agents, Docs |
| Nemotron 3 Super 120B A12B | 5-bit | 256K | 30 tok/s | Chat, Coding, Agents, Docs |
| Nemotron 3.5 Bolt 74B A9B | 8-bit | 256K | 28 tok/s | Chat, Coding, Agents, Docs |
| Qwen3.6 35B A3B | 8-bit | 256K | 74 tok/s | Chat, Coding, Agents, Docs |
| Gemma 4 31B | 8-bit | 256K | 12 tok/s | Chat, Coding, Agents, Docs |
| Gemma 4 26B A4B | 8-bit | 256K | 71 tok/s | Chat, Coding, Agents, Docs |
| GLM-4.7 Flash | 8-bit | 198K | 70 tok/s | Chat, Coding, Agents |
| Mistral Small 4 119B | 5-bit | 256K | 61 tok/s | Chat, Docs |
| gpt-oss 120B | MXFP4 | 128K | 72 tok/s | Chat, Coding |
| Qwen3 Coder Next 80B | 8-bit | 256K | 68 tok/s | Coding, Agents |
| Nemotron 3.5 Lightning 30B A3B | 8-bit | 256K | 66 tok/s | Agents, Docs |
| Devstral Small 2 24B | FP8 | 384K | 16 tok/s | Coding |
| LFM2.5 24B A2B | 8-bit | 128K | 93 tok/s | Agents |
| Mistral Small 3.2 24B | 8-bit | 128K | 16 tok/s | Chat, Coding, Agents |
| Llama 3.3 70B | 8-bit | 64K | 5.4 tok/s | Chat, Coding, Agents, Docs |
| Olmo 3.1 32B | 8-bit | 64K | 11 tok/s | Chat, Coding, Agents, Docs |
| Qwen3 32B | 8-bit | 40K | 11 tok/s | Not rated |
| Qwen3.6 27B | 8-bit | 256K | 13 tok/s | Text and images |
| Qwen3 Coder 30B A3B Instruct | 8-bit | 256K | 69 tok/s | Coding |
| Qwen3.5 27B | 8-bit | 256K | 13 tok/s | Text and images |
| Qwen3.5 35B A3B | 8-bit | 256K | 74 tok/s | Text and images |
| NVIDIA Nemotron 3 Nano 30B A3B BF16 | 8-bit | 256K | 66 tok/s | Not rated |
| GLM 4.5 Air | 6-bit | 32K | 27 tok/s | Not rated |
| Inkling Small | 2-bit | 4K | 61 tok/s | Text and images |
| Qwen3 30B A3B | 8-bit | 40K | 69 tok/s | Not rated |
| MiniMax M2.5 | 2-bit | 40K | 52 tok/s | Not rated |
| Qwen3 30B A3B Instruct 2507 | 8-bit | 256K | 69 tok/s | Not rated |
| Laguna XS.2 | 8-bit | 256K | 74 tok/s | Not rated |
| OTel 2.0 LLM 31B IT | 8-bit | 256K | 12 tok/s | Not rated |
| Laguna M.1 | 2-bit | 40K | 32 tok/s | Not rated |
| Qwen2.5 32B Instruct | 8-bit | 32K | 11 tok/s | Not rated |
| MiniCPM V 4.7 35B A3B | 8-bit | 256K | 74 tok/s | Text and images |
| Laguna XS 2.1 | 8-bit | 256K | 74 tok/s | Not rated |
| Diffusiongemma 26B A4B IT | 8-bit | 256K | 69 tok/s | Text and images |
| Qwen2.5 VL 32B Instruct | 8-bit | 125K | 11 tok/s | Text and images |
| Qwen2.5 Coder 32B Instruct | 8-bit | 32K | 11 tok/s | Coding |
| Laguna S 2.1 | 6-bit | 16K | 42 tok/s | Not rated |
| Qwen3 VL 30B A3B Instruct | 8-bit | 256K | 69 tok/s | Text and images |
| DeepSeek R1 Distill Qwen 32B | 8-bit | 128K | 11 tok/s | Not rated |
| Gemma 3 27B IT | 8-bit | 128K | 14 tok/s | Text and images |
| Qwen3 VL 235B A22B Instruct | 2-bit | 44K | 30 tok/s | Text and images |
| Qwen3 235B A22B | 2-bit | 40K | 30 tok/s | Not rated |
| Ornith 1.5 35B A3B | 8-bit | 256K | 74 tok/s | Not rated |
| Ornith 1.5 122B | 5-bit | 256K | 34 tok/s | Not rated |
| Ornith 1.5 397B | 1.6-bit | 256K | 47 tok/s | Not rated |
| Qwen3 VL 32B Instruct | 8-bit | 224K | 11 tok/s | Text and images |
| Step 3.5 Flash | 2-bit | 52K | 67 tok/s | Not rated |
| Qwen2.5 72B Instruct | 8-bit | 32K | 5.2 tok/s | Not rated |
| Qwen3 Next 80B A3B Instruct | 8-bit | 256K | 68 tok/s | Not rated |
| Llama 4 Scout 17B 16E Instruct | 6-bit | 36K | 23 tok/s | Text and images |
| MiniMax M2 | 2-bit | 40K | 52 tok/s | Not rated |
| Qwen2.5 VL 72B Instruct | 8-bit | 56K | 5.2 tok/s | Text and images |
| Qwen1.5 7B | 8-bit | 32K | 48 tok/s | Not rated |
Best fit is the highest quality, up to 8-bit, that leaves the use case’s context. Ranked by our use-case scores, then by Hugging Face downloads. Speeds are rough, plus or minus 25%. The top 60 of 119 are shown.
Image and video models
| Model | Best fit | Per output | Strongest at |
|---|---|---|---|
| MiniMax H3 | 16-bit | 62 min | Video |
| LTX-2.5 | 16-bit | 6.5 min | Video |
| Qwen-Image | 16-bit | 37 min | Images |
| FLUX.2 [dev] | 16-bit | 17 min | Images |
| Wan2.2 I2V A14B | 16-bit | 5.5 h | Video |
| Wan2.1 I2V 14B | 16-bit | 5.5 h | Video |
| Wan 2.1 T2V-14B | 16-bit | 6.9 h | Video |
| Qwen Image 2512 | 16-bit | 37 min | Images |
| Wan 2.2 A14B | 16-bit | 5.5 h | Video |
Only just too big
- GLM-5.3 Flash
- MiMo V2.6 Flash
- DeepSeek V4 Flash
- Hy3
- Nemotron 3 Ultra 550B A55B
- MiMo V2.5
- GLM 5.3 Flash BF16
- Hy3 Preview
- Qwen3 Coder 480B A35B Instruct
- MiMo V2 Flash
- Beam 1.0 501B A23B
These need up to a quarter more memory than the M4 Max 128 GB has. Spilling into system RAM or a lower precision can get them running, slowly.
Questions
What is the best LLM for the M4 Max 128 GB?
Of the open models we track, the strongest that run well on the M4 Max 128 GB are Qwen3.8 Flash-Next, MiniMax M3, Qwen3.8 27B. 119 text models fit it in all.
Calculated from published specs, not benchmarked. All machines