- Can I run it locally?
- Other open models
Phi, Nemotron, Step and LFM VRAM requirements: can I run them locally?
Models from labs with one or two popular downloads share this page. A new model from a lab without its own page shows up here the day it passes our download line.
Check your machine
Won't fit 125 GB
Weights, cache and overhead add up to more than the card and the RAM together.
- Weights 122 GB
- Context cache 1.7 GB
- Overhead 0.8 GB
- Fast memory
- With system RAM
What we found
The architectures here vary more than on any family page. Nemotron 3.5 Lightning 30B A3B has attention in only 6 of its 52 layers. Another 23 are Mamba layers, which keep a fixed-size state, so a million-token context costs 6.5 GB of cache.
Phi 4 mini goes the other way. Its 4-bit file is 2.5 GB, and a 32K context adds 4.3 GB of cache, more than the model itself.
Step 3.7 Flash from StepFun is the big one, 125 GB at 4-bit with a 32K context. A DGX Spark holds it, and so does a 128 GB Mac with its GPU limit raised.
Step 3.7 Flash
197 billion parameters · mixture of experts, 8 of 288 active · up to 256K context · 308 thousand downloads in 30 days
Step 3.7 Flash needs about 125 GB at UD-Q4_K_M with a 32K-token context. The file itself is 122 GB, and the cache for 32K tokens adds 1.7 GB. No single consumer card holds it. A Mac with 128 GB of unified memory does once you raise the GPU memory limit, and a PC can split it between a GPU and system RAM. It's a mixture-of-experts model (8 of 288 experts per token), which is why people run it with the experts parked in system RAM and the rest on a 24 GB card.
33 of its layers only look back 512 tokens, so they stop growing early, and the whole cache at 256K is 13.0 GB.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-IQ1_M | 56.8 GB | 58.1 GB | 59.3 GB | 96 GB | 64 GB* |
UD-Q2_K_XL | 65.8 GB | 67.1 GB | 68.3 GB | 96 GB | 96 GB |
UD-Q3_K_XL | 89.4 GB | 90.7 GB | 91.9 GB | 96 GB | 96 GB* |
UD-Q4_K_M | 122 GB | 123 GB | 125 GB | none | 128 GB* |
UD-Q5_K_XL | 146 GB | 148 GB | 149 GB | none | 256 GB |
UD-Q6_K_XL | 179 GB | 180 GB | 181 GB | none | 256 GB |
Q8_0 | 209 GB | 211 GB | 212 GB | none | 256 GB* |
BF16 | 394 GB | 395 GB | 397 GB | none | 512 GB |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
Phi 4 mini
3.8 billion parameters · up to 128K context · 119 thousand downloads in 30 days
At Q4_K_M, the Phi 4 mini file is 2.5 GB, and with 32K tokens of context and runtime overhead it comes to about 7.7 GB. Even a 8 GB card (RTX 4060, RTX 3070) holds that. On a Mac, 16 GB of unified memory is enough.
The cache is the part that grows: 128 KB a token, or 17.2 GB if you use all 128K tokens it supports.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
Q2_K | 1.7 GB | 3.7 GB | 6.9 GB | 8 GB | 16 GB |
Q3_K_M | 2.1 GB | 4.1 GB | 7.3 GB | 8 GB | 16 GB |
Q4_K_M | 2.5 GB | 4.5 GB | 7.7 GB | 8 GB | 16 GB |
Q5_K_M | 2.8 GB | 4.8 GB | 8.1 GB | 8 GB | 16 GB |
Q6_K | 3.2 GB | 5.1 GB | 8.4 GB | 8 GB | 16 GB |
Q8_0 | 4.1 GB | 6.1 GB | 9.3 GB | 12 GB | 16 GB |
BF16 | 7.7 GB | 9.7 GB | 12.9 GB | 16 GB | 24 GB |
Smallest GPU and Mac are for the 32K total.
Sources: GGUF files · config.json
Nemotron 3.5 Lightning 30B A3B
32.9 billion parameters · mixture of experts, 6 of 128 active · up to 1M context · 68 thousand downloads in 30 days
Nemotron 3.5 Lightning 30B A3B needs about 26.3 GB at UD-Q4_K_M with a 32K-token context. The file itself is 25.3 GB, and the cache for 32K tokens adds 0.3 GB. That fits a RTX 5090 (32 GB). It's a mixture-of-experts model (6 of 128 experts per token), which is why people run it with the experts parked in system RAM and the rest on a 24 GB card.
Only 6 of its 52 layers keep a growing cache (the rest carry a fixed-size state), so the full 1M context costs 6.5 GB of cache.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-IQ1_M | 19.4 GB | 20.3 GB | 20.4 GB | 24 GB | 24 GB* |
UD-IQ2_M | 19.4 GB | 20.3 GB | 20.5 GB | 24 GB | 24 GB* |
UD-Q3_K_XL | 21.2 GB | 22.1 GB | 22.3 GB | 24 GB | 32 GB |
UD-Q4_K_M | 25.3 GB | 26.1 GB | 26.3 GB | 32 GB | 32 GB* |
UD-Q5_K_XL | 30.4 GB | 31.3 GB | 31.4 GB | 32 GB | 36 GB* |
UD-Q6_K_XL | 35.0 GB | 35.9 GB | 36.0 GB | 48 GB | 48 GB |
Q8_0 | 35.0 GB | 35.9 GB | 36.0 GB | 48 GB | 48 GB |
BF16 | 65.9 GB | 66.7 GB | 66.9 GB | 96 GB | 96 GB |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
LFM2.5 1.2B
1.2 billion parameters · up to 125K context · 42 thousand downloads in 30 days
At Q4_K_M, the LFM2.5 1.2B file is 0.7 GB, and with 32K tokens of context and runtime overhead it comes to about 1.8 GB. Even a 8 GB card (RTX 4060, RTX 3070) holds that. On a Mac, 16 GB of unified memory is enough.
Most of its layers use linear attention, which keeps a fixed state instead of a cache, so the cache grows by about 12 KB a token and reaches 1.6 GB at 125K.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-Q2_K_XL | 0.5 GB | 1.2 GB | 1.5 GB | 8 GB | 16 GB |
UD-Q3_K_XL | 0.6 GB | 1.3 GB | 1.6 GB | 8 GB | 16 GB |
Q4_K_M | 0.7 GB | 1.5 GB | 1.8 GB | 8 GB | 16 GB |
Q5_K_M | 0.8 GB | 1.6 GB | 1.9 GB | 8 GB | 16 GB |
Q6_K | 1.0 GB | 1.7 GB | 2.0 GB | 8 GB | 16 GB |
Q8_0 | 1.2 GB | 2.0 GB | 2.3 GB | 8 GB | 16 GB |
BF16 | 2.3 GB | 3.1 GB | 3.4 GB | 8 GB | 16 GB |
Smallest GPU and Mac are for the 32K total.
Sources: GGUF files · config.json
Ornith 1.0 9B
9.0 billion parameters · up to 256K context · 483 thousand downloads in 30 days
Ornith 1.0 9B needs about 7.8 GB at Q4_K_M with a 32K-token context. The file itself is 5.7 GB, and the cache for 32K tokens adds 1.1 GB. That fits a 8 GB card (RTX 4060, RTX 3070). On a Mac, 16 GB of unified memory is enough.
Only 8 of its 32 layers keep a growing cache (the rest carry a fixed-size state), so the full 256K context costs 8.6 GB of cache.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-Q2_K_XL | 4.3 GB | 5.7 GB | 6.5 GB | 8 GB | 16 GB |
UD-Q3_K_XL | 5.1 GB | 6.4 GB | 7.2 GB | 8 GB | 16 GB |
Q4_K_M | 5.7 GB | 7.0 GB | 7.8 GB | 8 GB | 16 GB |
Q5_K_M | 6.5 GB | 7.9 GB | 8.7 GB | 12 GB | 16 GB |
Q6_K | 7.5 GB | 8.8 GB | 9.6 GB | 12 GB | 16 GB |
Q8_0 | 9.5 GB | 10.9 GB | 11.7 GB | 12 GB | 16 GB* |
BF16 | 17.9 GB | 19.3 GB | 20.1 GB | 24 GB | 24 GB* |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
North Mini Code 1.0
30.5 billion parameters · mixture of experts, 8 of 128 active · up to 488K context · 316 thousand downloads in 30 days
At UD-Q4_K_M, the North Mini Code 1.0 file is 19.2 GB, and with 32K tokens of context and runtime overhead it comes to about 21.5 GB. The smallest card that holds it whole is a 24 GB card (RTX 3090, RTX 4090). It routes each token through 8 of 128 experts, so it runs faster than its size suggests.
A sliding window of 4,096 tokens on 36 layers keeps the cache small: 13.7 GB at the full 488K.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-IQ1_M | 9.4 GB | 11.0 GB | 11.6 GB | 12 GB | 16 GB* |
UD-Q2_K_XL | 10.5 GB | 12.1 GB | 12.7 GB | 12 GB | 16 GB* |
UD-Q3_K_XL | 14.3 GB | 15.9 GB | 16.6 GB | 16 GB | 24 GB |
UD-Q4_K_M | 19.2 GB | 20.8 GB | 21.5 GB | 24 GB | 24 GB* |
UD-Q5_K_XL | 23.0 GB | 24.6 GB | 25.2 GB | 24 GB | 32 GB* |
UD-Q6_K_XL | 27.9 GB | 29.4 GB | 30.1 GB | 32 GB | 36 GB* |
Q8_0 | 32.4 GB | 34.0 GB | 34.7 GB | 48 GB | 48 GB |
BF16 | 61.0 GB | 62.6 GB | 63.3 GB | 96 GB | 64 GB* |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
Ornith 1.0 35B
34.7 billion parameters · mixture of experts, 8 of 256 active · up to 256K context · 24 thousand downloads in 30 days
Ornith 1.0 35B needs about 23.9 GB at UD-Q4_K_M with a 32K-token context. The file itself is 22.1 GB, and the cache for 32K tokens adds 0.7 GB. That fits a 24 GB card (RTX 3090, RTX 4090). It routes each token through 8 of 256 experts, so it runs faster than its size suggests.
Only 10 of its 40 layers keep a growing cache (the rest carry a fixed-size state), so the full 256K context costs 5.4 GB of cache.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-IQ1_S | 10.5 GB | 11.8 GB | 12.3 GB | 12 GB | 16 GB* |
UD-Q2_K_XL | 12.3 GB | 13.5 GB | 14.0 GB | 16 GB | 24 GB |
UD-Q3_K_XL | 16.8 GB | 18.0 GB | 18.5 GB | 24 GB | 24 GB* |
UD-Q4_K_M | 22.1 GB | 23.4 GB | 23.9 GB | 24 GB | 32 GB* |
UD-Q5_K_XL | 26.5 GB | 27.8 GB | 28.3 GB | 32 GB | 32 GB* |
UD-Q6_K_XL | 31.8 GB | 33.1 GB | 33.6 GB | 32 GB | 36 GB* |
Q8_0 | 36.9 GB | 38.1 GB | 38.6 GB | 48 GB | 48 GB |
BF16 | 69.4 GB | 70.6 GB | 71.1 GB | 96 GB | 96 GB |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
LFM2.5 8B A1B
8.5 billion parameters · mixture of experts, 4 of 32 active · up to 125K context · 21 thousand downloads in 30 days
At UD-Q4_K_M, the LFM2.5 8B A1B file is 5.3 GB, and with 32K tokens of context and runtime overhead it comes to about 6.5 GB. Even a 8 GB card (RTX 4060, RTX 3070) holds that. On a Mac, 16 GB of unified memory is enough. It routes each token through 4 of 32 experts, so it runs faster than its size suggests.
Most of its layers use linear attention, which keeps a fixed state instead of a cache, so the cache grows by about 12 KB a token and reaches 1.6 GB at 125K.
| Quant | File | At 8K | At 32K | Smallest GPU | Smallest Mac |
|---|---|---|---|---|---|
UD-IQ1_M | 2.6 GB | 3.4 GB | 3.7 GB | 8 GB | 16 GB |
UD-Q2_K_XL | 2.9 GB | 3.8 GB | 4.1 GB | 8 GB | 16 GB |
UD-Q3_K_XL | 4.0 GB | 4.9 GB | 5.2 GB | 8 GB | 16 GB |
UD-Q4_K_M | 5.3 GB | 6.2 GB | 6.5 GB | 8 GB | 16 GB |
UD-Q5_K_XL | 6.4 GB | 7.3 GB | 7.6 GB | 8 GB | 16 GB |
UD-Q6_K_XL | 7.7 GB | 8.6 GB | 8.9 GB | 12 GB | 16 GB |
Q8_0 | 9.0 GB | 9.9 GB | 10.2 GB | 12 GB | 16 GB |
BF16 | 16.9 GB | 17.8 GB | 18.1 GB | 24 GB | 24 GB* |
Smallest GPU and Mac are for the 32K total. * with the Mac's GPU memory limit raised.
Sources: GGUF files · config.json
How the numbers are worked out
Total memory is three parts added together. Weights are the exact size of the GGUF file on Hugging Face, all parts summed when a quant is split. The context cache (the KV cache) comes from each model's own config.json:
cache = layers × kv_heads × head_dim × 2 (K and V) × 2 bytes × tokens
Sliding-window layers stop at their window plus llama.cpp's 512-token batch. Linear-attention and Mamba layers keep a fixed state instead, which we add once. Latent-attention models (DeepSeek, Kimi, GLM-5) cache kv_lora_rank + rope dimensions per layer, the way llama.cpp does. Overhead is our allowance: 0.5 GB for the GPU context plus the output buffer, 512 × vocabulary × 4 bytes.
A "24 GB" card holds 24 GiB, which is 25.8 GB in the decimal units file sizes use, and we count it that way. Macs get the share Metal allows by default, two thirds of unified memory up to 32 GB and three quarters above (llama.cpp discussion 2182); CPU-only machines and the DGX Spark keep 4 GB back for the operating system. All of it assumes llama.cpp's defaults: an f16 cache and one conversation at a time. Quantizing the cache to q8_0 halves it.
Files are Unsloth's GGUF uploads, the most downloaded quantizations on Hugging Face, and the list refreshes daily from Hugging Face's API. A model joins when its repo passes 20,000 downloads in 30 days. Last checked September 25, 2026.