Sage-A1.5
Sage-A1.5 loads in 21.7 GB at Q4_K_M and decodes at the speed of its 3B active path, comfortably from 32 GB of memory, Macs included, starting at the MacBook Pro M5 Pro 48GB. Below is the memory math we ran for it, the smallest card that fits, and the fastest machine we track.
- Parameters
- 35.95B(3B active)
- Format
- Q4_K_M
- Min memory
- 32 GB
- Best for
- Scientific research agents, Long-horizon tasks
How much memory does Sage-A1.5 need?
Sage-A1.5 is a 35.95B mixture-of-experts model (3B active per token) built for scientific research agents, long-horizon tasks. The weights alone take 21.7 GB at Q4_K_M; add context and the OS, and it belongs in the 32 GB memory class. Every number below comes from the ModelFit engine: speeds are estimates, labeled est., because we do not benchmark hardware ourselves.
Run with llama.cpp: llama-server -hf InternScience/Sage-A1.5-Q4_K_M-GGUF:Q4_K_M.
KV = fp16 estimate (q8_0 cache roughly halves it). "Comfortable" = weights + KV within the engine's tiered budget (~70-85% of memory).
Which machines run Sage-A1.5?
- Smallest GPU
- NVIDIA GeForce RTX 5090est. 346 tok/s · 32 GB VRAMtypical error 15%
- Fastest
- NVIDIA GeForce RTX 5090est. 346 tok/stypical error 15%
- Runs on a Mac?
- MacBook Pro M5 Pro 48GBest. 99 tok/stypical error 10%
Verdicts come from the same engine that powers our fit checker: memory capacity first, bandwidth for speed. How we compute these numbers · What typical error means · Hugging Face model card
Get told when a better model fits your MacBook Air M5 16 GB
Get one email when a new open-weight model is a better fit for this machine.
For: MacBook Air M5 16 GB
You'll also get the weekly ModelFit email. By signing up you agree to our Privacy Policy. Unsubscribe from either email anytime, in one click.
Frequently asked questions
How much memory does Sage-A1.5 need?
21.7 GB for the Q4_K_M weights, plus ~0.3 GB of KV-cache at 16k context, about 22.0 GB total. Comfortable from 32 GB of VRAM or unified memory.
Does Sage-A1.5 run on a Mac?
Yes, from the MacBook Pro M5 Pro 48GB (est. 99 tok/s, typical error 10%). Unified memory means the RAM budget is the only limit.
What is the smallest GPU for Sage-A1.5?
The NVIDIA GeForce RTX 5090 (32 GB VRAM) is the smallest tracked card that runs Sage-A1.5 comfortably, est. 346 tok/s, typical error 15%.
Cite this page
ModelFit: Sage-A1.5, specs, memory math and hardware verdicts. https://modelfit.io/models/sage-a1.5 (dataset updated 2026-10-30, CC BY 4.0).