QuantDiff

Best Sage-A1.5 quant for a 24GB GPU

Short answer: IQ4_XS-exps from na0x2c6. Of the 4 Sage-A1.5 files we tested that fit fully in 24GB with a 8192-token context, it has the highest measured accuracy (90.2% on GSM8K + MMLU + ARC-C, 95% CI 87.3–92.5). Accuracy and speed only, no KLD: reference too large.

Smallest diff that fits 24GB
IQ4_XS-exps by na0x2c6
File size
19.3 GB
no KLD: reference too large
Accuracy (500 items)
90.2% (87.3–92.5)
87.0%
Accuracy
Generation ()
50 t/s on Apple M2 Pro 32GB full GPU
Sage-A1.5-IQ4_XS-exps.gguf
How to run this file

From na0x2c6/Sage-A1.5-GGUF, file Sage-A1.5-IQ4_XS-exps.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/na0x2c6/Sage-A1.5-GGUF:Sage-A1.5-IQ4_XS-exps.gguf

: open the Discover tab, search na0x2c6/Sage-A1.5-GGUF, pick the IQ4_XS-exps download, then load it from the Chat tab.

Step-by-step guide

Runner-up
Q4_K_M by InternScience
File size
21.7 GB
no KLD: reference too large
Accuracy (500 items)
89.8% (86.8–92.2)
85.0%
Accuracy
Generation ()
51 t/s on Apple M2 Pro 32GB full GPU
Sage-A1.5-Q4_K_M.gguf
How to run this file

From InternScience/Sage-A1.5-Q4_K_M-GGUF, file Sage-A1.5-Q4_K_M.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/InternScience/Sage-A1.5-Q4_K_M-GGUF:Sage-A1.5-Q4_K_M.gguf

: open the Discover tab, search InternScience/Sage-A1.5-Q4_K_M-GGUF, pick the Q4_K_M download, then load it from the Chat tab.

Step-by-step guide

Fastest that fits
Q4_K_M by InternScience
File size
21.7 GB
no KLD: reference too large
Accuracy (500 items)
89.8% (86.8–92.2)
85.0%
Accuracy
Generation ()
51 t/s on Apple M2 Pro 32GB full GPU
Sage-A1.5-Q4_K_M.gguf
How to run this file

From InternScience/Sage-A1.5-Q4_K_M-GGUF, file Sage-A1.5-Q4_K_M.gguf.

(downloads it, then starts a chat; type /bye to quit):

ollama run hf.co/InternScience/Sage-A1.5-Q4_K_M-GGUF:Sage-A1.5-Q4_K_M.gguf

: open the Discover tab, search InternScience/Sage-A1.5-Q4_K_M-GGUF, pick the Q4_K_M download, then load it from the Chat tab.

Step-by-step guide

Other budgets: 8GB12GB16GB · all 4 Sage-A1.5 quants

How we picked

We estimate VRAM need as file size + f16 KV cache for 8192 tokens + about 0.6GB of compute buffers, and keep 0.5GB free for your desktop. Among files that fit, we rank by mean KL divergence against the BF16 model's logits, then by tool-call score. If you need a longer context, step down one quant level or quantize the KV cache (-ctk q8_0 -ctv q8_0). Methodology.

GPU too small? Rent one.

Run a 32B model at Q6/Q8, or a 70B, on a cloud GPU for an hour or two instead of buying hardware.

RunPod

Cloud GPU pods with 24-80GB cards (RTX 4090 / 5090 / A100 / H100) for big quants or BF16 reference runs.

Rent on RunPod
Vast.ai

GPU marketplace for renting a 24-48GB card for an evening of testing.

Rent on Vast.ai
Your GPU cloud here

Sponsor slot. Sponsors never influence scores or rankings.

Sponsor QuantDiff

Referral links: QuantDiff may earn credit if you sign up. Referral partners have no say in scores or rankings.