Models › Orion-26B-A4B-v1.1

Orion-26B-A4B-v1.1: VRAM needs and speed by GPU

At Q4_K_M, Orion-26B-A4B-v1.1 needs about 19.4 GB of VRAM. No measured runs yet. Speeds below are estimates. Submit a run.

25.23B parameters (mixture-of-experts, ~3.9B active) · 262,144-token context · GGUF published Sep 13, 2026 · GGUF files on Hugging Face

Memory needed per quantization

QuantFile sizeNeeds (estimate)Smallest card it fits
IQ2_XS9.9 GB11.3 GB12 GB cards
IQ2_S10.3 GB11.7 GB16 GB cards
IQ2_M10.9 GB12.3 GB16 GB cards
Q2_K11.6 GB13.0 GB16 GB cards
Q3_K_S12.4 GB13.8 GB16 GB cards
IQ3_XXS12.6 GB14.0 GB16 GB cards
IQ3_XS12.8 GB14.3 GB16 GB cards
Q3_K_M13.8 GB15.2 GB16 GB cards
IQ4_XS14.3 GB15.7 GB24 GB cards
IQ3_M14.3 GB15.8 GB24 GB cards
Q4_014.8 GB16.2 GB24 GB cards
Q3_K_L14.8 GB16.2 GB24 GB cards
IQ4_NL15.6 GB17.1 GB24 GB cards
Q4_K_S16.1 GB17.5 GB24 GB cards
Q4_116.1 GB17.6 GB24 GB cards
Q4_K_M18.0 GB19.4 GB24 GB cards
Q4_K_L18.2 GB19.6 GB24 GB cards
Q5_K_S18.4 GB19.8 GB24 GB cards
Q5_K_M20.0 GB21.4 GB24 GB cards
Q6_K_S23.1 GB24.6 GB32 GB cards
Q6_K23.8 GB25.2 GB32 GB cards
Q6_K_L24.5 GB26.0 GB32 GB cards
Q8_026.9 GB28.3 GB32 GB cards
BF1650.5 GB51.9 GB96 GB cards

Expected speed on popular hardware

HardwareBest quant that fitsExpected tok/s
NVIDIA RTX 3060IQ2_XS101 estimate
NVIDIA RTX 4060 TiDoes not fit
NVIDIA RTX 3090Q5_K_M107 estimate
NVIDIA RTX 4090Q5_K_M118 estimate
NVIDIA RTX 5090Q8_0139 estimate
AMD Radeon RX 7900 XTXQ5_K_M119 estimate
Apple M4 Pro (20-core GPU) (64 GB)Q8_028.0 estimate
Apple M3 Max (40-core GPU) (64 GB)Q8_040.9 estimate
AMD Ryzen AI Max+ 395 (Radeon 8060S)Q8_025.2 estimate

Check your own card

Too big for your card?

Run Orion-26B-A4B-v1.1 on a rented 24-80 GB GPU by the hour.

Rent a bigger GPU on RunPod

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.