Can the Mac 128 GB run LTX-2.5 (22B)?
Yes. 16-bit fits on the Mac 128 GB with 68.6 GB to spare.
16-bit (42.0 GB) needs about 46.8 GB in use, so 68.6 GB of the 115.4 GB stays free: room for a larger image, a LoRA or a ControlNet.
01What to download
| Part | File | Folder | Size | |
|---|---|---|---|---|
| Model | ltx-2.5-22b-dev-transformer-bf16.safetensors 16-bit · Lightricks/LTX-2.5 | models/diffusion_models | 42.0 GB | Download → |
| Text encoder | gemma4-12b-with-proj-ltx-2.5-bf16.safetensors Gemma 4 12B + LTX projection BF16 · Lightricks/LTX-2.5 | models/text_encoders | 26.3 GB | Download → |
| VAE | ltx-2.5-video-vae-bf16.safetensors LTX-2.5 video VAE (DiffVAE) BF16 · Lightricks/LTX-2.5 | models/vae | 1.5 GB | Download → |
| Also needed | ltx-2.5-audio-vae-bf16.safetensors LTX-2.5 audio VAE + vocoder BF16 · Lightricks/LTX-2.5 | models/vae | 0.4 GB | Download → |
| Total download · keep about the same free on disk | 70.1 GB | |||
Only for some workflows: LTX-2.5 latent spatial upscaler x2 (1.0 GB, official two-stage template); Gemma 4 E2B INT8 (prompt enhancer, optional) (5.2 GB, only if prompt enhancer enabled). On a Mac, FP8 text encoders do not load, so a 16-bit one is listed. Sizes from Hugging Face, 2026-10-01.
Shared memory: about 76.1 GB of your 128 GB. On this machine the model, the text encoder, the VAE and the operating system all use the same memory. That leaves room, so nothing has to be swapped to disk. calculated
02Every LTX-2.5 file on 115.4 GB
| File | Size | Needed | On this card | Quality | Download |
|---|---|---|---|---|---|
| 16-bit ← SAFETENSORS · Lightricks | 42.0 GB | 46.8 GB | fits · 68.6 GB spare | the original weights | Hugging Face → |
| Q8_0 GGUF · realrebelai | 23.6 GB | 28.4 GB | fits · 87.0 GB spare | practically identical to the original | Hugging Face → |
| Q6_K GGUF · realrebelai | 18.7 GB | 23.5 GB | fits · 91.9 GB spare | very close to the original | Hugging Face → |
| Q5_K_M GGUF · realrebelai | 16.8 GB | 21.6 GB | fits · 93.8 GB spare | close; small differences in fine detail | Hugging Face → |
| Q4_K_M GGUF · realrebelai | 15.1 GB | 19.9 GB | fits · 95.5 GB spare | good; some loss in fine detail and text | Hugging Face → |
| Q3_K_M GGUF · realrebelai | 11.5 GB | 16.3 GB | fits · 99.1 GB spare | noticeable loss of detail | Hugging Face → |
| Q2_K GGUF · realrebelai | 8.8 GB | 13.6 GB | fits · 101.8 GB spare | heavy loss; a last resort | Hugging Face → |
“Needed” adds 4 GB of working memory for LTX-2.5 and a 0.8 GB reserve. Sizes 2026-10-01.
03Measured and reported results
Nobody has sent measured numbers for this pair yet. If you run LTX-2.5 on a Mac 128 GB, send your time per image and peak VRAM and it will appear here, credited.
04About the Mac 128 GB
A Mac with 128 GB shares that memory between the CPU and the GPU. On current macOS the GPU may use about 115.4 GB of it by default (older macOS versions: about 103.1 GB); that is the figure used here. ComfyUI can go past it, but macOS then starts compressing and swapping memory and everything slows down. ComfyUI runs on Apple GPUs through PyTorch's MPS backend. 16-bit and GGUF files work; FP8 and INT8 files do not save memory on a Mac, so they are skipped. Speed is the catch: even the fastest Macs are several times slower per image than a desktop RTX card. Chips sold with 128 GB: M1 Ultra, M2 Ultra, M3 Max, M4 Max, M5 Max (memory bandwidth 300–800 GB/s - the higher, the faster). Everything about Macs and local AI →
128 GB unified · ~115.4 GB for the GPU · 300–800 GB/s · Apple Silicon. Everything that runs on it →
05Questions
How much VRAM does LTX-2.5 need?
Around 13.6 GB with the smallest file (Q2_K) and 28.4 GB with an 8-bit file (Q8_0). The full 16-bit file needs about 46.8 GB.
Which LTX-2.5 file should I download for the Mac 128 GB?
16-bit (42.0 GB) from Lightricks/LTX-2.5.
How much do I need to download for LTX-2.5 on the Mac 128 GB?
70.1 GB: ltx-2.5-22b-dev-transformer-bf16.safetensors, gemma4-12b-with-proj-ltx-2.5-bf16.safetensors, ltx-2.5-video-vae-bf16.safetensors, ltx-2.5-audio-vae-bf16.safetensors.