RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run Axion 1.0 35B locally?

Axion 1.0 35B by Axion1-0 has 35B parameters (3.9B active per token, 256 experts, 8 used per token). At 4-bit with an 8K context it needs about 20.7 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track. See all models that fit in 24 GB of VRAM.

20.7 GBQ4_K_M · 8K context
RTX 5090smallest consumer GPU that fits
M3 Pro 36GBsmallest Mac that fits
256Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from prithivMLmods/Axion-1.0-35B-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K)rough13.2 GB13.3 GB13.8 GB15.6 GB
3-bit (Q3_K_M) •usable16.6 GB16.6 GB17.1 GB19.0 GB
4-bit (Q4_K_M) •good, most popular20.7 GB20.7 GB21.2 GB23.1 GB
5-bit (Q5_K_M) •very good24.0 GB24.1 GB24.5 GB26.4 GB
6-bit (Q6_K) •near lossless27.5 GB27.6 GB28.1 GB29.9 GB
8-bit (Q8_0)effectively lossless35.7 GB35.8 GB36.2 GB38.1 GB
16-bit (F16/BF16) •original weights65.6 GB65.6 GB66.1 GB68.0 GB

Which GPUs and Macs run Axion 1.0 35B?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBNo———details
RTX 4060 Ti 16GB16 GBFitsQ2_K13.3 GB197details
RTX 309024 GBFitsQ3_K_M16.6 GB502details
RTX 409024 GBFitsQ3_K_M16.6 GB541details
RTX 509032 GBFitsQ5_K_M24.1 GB652details
RTX 5070 Ti16 GBFitsQ2_K13.3 GB612details
RX 7900 XTX24 GBFitsQ3_K_M16.6 GB515details
Mac M4 Pro 48GB48 GBFitsQ6_K27.6 GB86details
Mac M4 Max 128GB128 GBFitsF1665.6 GB71details
Mac M3 Ultra 512GB512 GBFitsF1665.6 GB106details
DGX Spark128 GBFitsF1665.6 GB35details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBFitsQ2_K13.3 GB306details
RTX 40608 GBNo———details
Arc B58012 GBNo———details
RTX 3060 Ti8 GBNo———details
RTX 30708 GBNo———details
RTX 3080 10GB10 GBNo———details
RTX 3080 12GB12 GBNo———details
RTX 3090 Ti24 GBFitsQ3_K_M16.6 GB541details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBFitsQ2_K13.3 GB459details
RTX 408016 GBFitsQ2_K13.3 GB490details
RTX 4080 Super16 GBFitsQ2_K13.3 GB503details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBFitsQ2_K13.3 GB656details
2x RTX 30902×24 GBFitsQ8_035.8 GB226details
2x RTX 40902×24 GBFitsQ8_035.8 GB243details
2x RTX 50902×32 GBFitsQ8_035.8 GB432details
RTX A600048 GBFitsQ8_035.8 GB185details
RTX 6000 Ada48 GBFitsQ8_035.8 GB232details
RTX PRO 6000 Blackwell96 GBFitsF1665.6 GB232details
L40S48 GBFitsQ8_035.8 GB208details
A100 40GB40 GBFitsQ6_K27.6 GB491details
A100 80GB80 GBFitsF1665.6 GB264details
H100 PCIe80 GBFitsF1665.6 GB259details
H100 SXM80 GBFitsF1665.6 GB434details
H200141 GBFitsF1665.6 GB623details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBFitsQ2_K13.3 GB426details
RX 7900 GRE16 GBFitsQ2_K13.3 GB394details
RX 7900 XT20 GBFitsQ3_K_M16.6 GB429details
RX 907016 GBFitsQ2_K13.3 GB437details
RX 9070 XT16 GBFitsQ2_K13.3 GB437details
Ryzen AI Max+ 395 128GB128 GBFitsF1665.6 GB33details
Arc A770 16GB16 GBFitsQ2_K13.3 GB383details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K13.3 GB68details
Mac M3 24GB24 GBFitsQ2_K13.3 GB68details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ3_K_M16.6 GB64details
Mac M1 Pro 32GB32 GBFitsQ3_K_M16.6 GB107details
Mac M3 Pro 36GB36 GBFitsQ4_K_M20.7 GB64details
Mac M4 Pro 24GB24 GBFitsQ2_K13.3 GB187details
Mac M4 Pro 64GB64 GBFitsQ8_035.8 GB66details
Mac M1 Max 64GB64 GBFitsQ8_035.8 GB96details
Mac M2 Max 96GB96 GBFitsQ8_035.8 GB96details
Mac M3 Max 128GB128 GBFitsF1665.6 GB52details
Mac M4 Max 64GB64 GBFitsQ8_035.8 GB132details
Mac M2 Ultra 192GB192 GBFitsF1665.6 GB104details
Mac M3 Ultra 256GB256 GBFitsF1665.6 GB106details

Questions

How much VRAM does Axion 1.0 35B need?

About 20.7 GB at 4-bit (Q4_K_M) with an 8K context: 19.7 GB of weights (real GGUF file size), 0.2 GB of KV cache and 0.9 GB of runtime overhead. At 8-bit it needs about 35.8 GB.

What is the smallest GPU that runs Axion 1.0 35B?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 5090 (32 GB).

Can I run Axion 1.0 35B on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M3 Pro 36GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

Qwen AgentWorld 35B A3B

35B · 21.6 GB at 4-bit

Yi 34B 200K

34B · 21.6 GB at 4-bit

Yi 34B

34B · 21.6 GB at 4-bit

Yi 34B Chat

34B · 21.6 GB at 4-bit

Yi 1.5 34B Chat

34B · 21.6 GB at 4-bit

Nous Hermes 2 Yi 34B

34B · 22.1 GB at 4-bit

Model facts

Hugging Face repoAxion1-0/Axion-1.0-35B
Parameters35,107,181,936 (35B) · ~3.9B active (computed from config)
Layers / hidden size40 / 2048
AttentionGQA (hybrid: 30 linear/state layers); 16 heads, 2 KV heads, head dim 256
KV cache per token (FP16)20.0 KiB
Max context262,144 tokens
Licenseapache-2.0
Released on HFOct 16, 2026
Downloads (30d) / likes1,863 / 76

Sources: Hugging Face model API · config.json (Axion1-0/Axion-1.0-35B) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.