Best Ornith 1.5 9B quant for a 24GB GPU

Short answer: Q5_K_M from ornith-ai. Of the 4 Ornith 1.5 9B files we tested that fit fully in 24GB with a 8192-token context, it has the highest measured accuracy (87.4% on GSM8K + MMLU + ARC-C, 95% CI 84.2–90.0). KLD pending: ranked by accuracy for now.

Smallest diff that fits 24GB

Q5_K_M by ornith-ai

File size: 6.19 GB
KLD: pending
Accuracy (500 items): 87.4% (84.2–90.0)
Tool calls: 89.0%
Accuracy: tied with best
Generation: 90 t/s full GPU

Ornith-1.5-9B-Q5_K_M.gguf

Runner-up

Q6_K by ornith-ai

File size: 7.04 GB
KLD: pending
Accuracy (500 items): 87.2% (84.0–89.8)
Tool calls: 89.0%
Accuracy: tied with best
Generation: 83 t/s full GPU

Ornith-1.5-9B-Q6_K.gguf

Fastest that fits

Q4_K_M by ornith-ai

File size: 5.38 GB
KLD: pending
Accuracy (500 items): 87.2% (84.0–89.8)
Tool calls: 87.5%
Accuracy: tied with best
Generation: 99 t/s full GPU

Ornith-1.5-9B-Q4_K_M.gguf

Other budgets: 8GB 12GB 16GB · all 4 Ornith 1.5 9B 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.