Carbon 3B — Hardware Requirements & GPU Compatibility
Chat
Carbon 3B is a 3.5B-parameter open language model from HuggingFaceBio. It supports a context window of up to 32,768 tokens. At Q4_K_M it needs about 2.47 GB of VRAM — see which GPUs and Macs can run it below.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 32K)Best on 24 GB (Q4_K_M · ~265 tok/s · 32K)Macs from 8 GB (Q4_K_M · ~29 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Carbon 3B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve HuggingFaceBio/Carbon-3B --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (8 GB headroom) · Q4_K_M
- Generation speed
- ~240 tok/s
- generation speed
- Cost per 1M output tokens
- $0.14
- per 1M output tokens
How Much VRAM Does Carbon 3B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 1.9 GB | 3.3 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 32K context · ~352 tok/s | |
| 2.1 GB | 3.5 GB | 3-bit medium — 3-bit medium quantization | Fits: 32K context · ~315 tok/s | |
| 2.5 GB | 3.9 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 32K context · ~265 tok/s | |
| 2.9 GB | 4.3 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 32K context · ~230 tok/s | |
| 3.2 GB | 4.7 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 32K context · ~202 tok/s | |
| 3.9 GB | 5.3 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 32K context · ~170 tok/s | |
| 7.3 GB | 8.7 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 32K context · ~90 tok/s |
Selected Q4_K_M
4-bit medium quantization3.1GBat 16K context
- Weights
- 2.1 GB
- KV cache
- 0.8 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 20.9 GB free · ~265 tok/s
≥50 tok/s needs ≥190 GB/s · RTX 5090 ~472, RTX 3090 Ti ~265
Which Hardware Can Run Carbon 3B?
Q4_K_M · 2.5 GBCarbon 3B (Q4_K_M) requires 2.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 4+ GB is recommended. Using the full 33K context window can add up to 1.4 GB, bringing total usage to 3.9 GB. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run Carbon 3B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 4.9 GB free · up to 32K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~200Simulate output on RTX 3080RTX 2080 Ti~162Simulate output on RTX 2080 TiRTX 3070 Ti~160Simulate output on RTX 3070 TiM3*8 GB~29Simulate output on iMac 24" M3 (8 GB)Orin Nano~27Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~19Simulate output on iPhone 1720 GB
fits · 16.9 GB free · up to 32K ctx
1 GPU
RX 7900 XT~194Simulate output on RX 7900 XT32–40 GB
fits · 28.9 GB free · up to 32K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~472Simulate output on RTX 5090M5 Max36 GB~130Simulate output on Mac Studio M5 Max (36 GB)M532 GB~44Simulate output on MacBook Air 13" M5 (32 GB)Orin~54Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~36Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Granite 4.0 MicroIBM | — | 3.7 GB | ~261 | same size |
| Deepseek Vl2 TinyDeepSeek | — | 9.4 GB | ~152 | same size |
| Llama 3.2 3B InstructMeta | — | 4.1 GB | ~309 | same size |
| Tiny Aya GlobalCohere | — | 9.4 GB | ~89 | same size |
| AI21 Jamba Reasoning 3BAI21 Labs | — | 2.3 GB | ~291 | same size |
| MiniMax M3 EAGLE3Inferact | — | 2.7 GB | ~284 | same size |
Specifications
- Publisher
- HuggingFaceBio
- Parameters
- 3.5B
- Architecture
- LlamaForCausalLM
- Context Length
- 32,768 tokens
- Vocabulary Size
- 155,776
- Release Date
- 2026-05-12
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Carbon 3B need?
Carbon 3B requires 2.5 GB of VRAM at Q4_K_M, or 7.3 GB at BF16. Full 33K context adds up to 1.4 GB (3.9 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 3.5B × 4.8 bits ÷ 8 = 2.1 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
KV Cache + Overhead ≈ 1.8 GB (at full 33K context)
VRAM usage by quantization
What's the best quantization for Carbon 3B?
For Carbon 3B, Q4_K_M (2.5 GB) offers the best balance of quality and VRAM usage. Q5_K_M (2.9 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 1.9 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Carbon 3B on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run Carbon 3B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 2.5 GB of usable unified memory (RAM minus macOS overhead).
Can I run Carbon 3B locally?
Yes — Carbon 3B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Carbon 3B?
At Q4_K_M, Carbon 3B can reach ~1943 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~265 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 2.47 × 0.65 = ~2105 tok/s
Estimated speed at Q4_K_M (2.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Carbon 3B?
At Q4_K_M, the download is about 2.07 GB. The full-precision BF16 version is 6.90 GB. The smallest option (Q2_K) is 1.47 GB.
Which GPUs can run Carbon 3B?
52 consumer GPUs can run Carbon 3B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run Carbon 3B?
163 devices with unified memory can run Carbon 3B at Q4_K_M (2.5 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.