Jan v3 4B Base Instruct — Hardware Requirements & GPU Compatibility
ChatCode
Jan v3 4B Base Instruct is a 4.4B-parameter open language model from janhq. It supports a context window of up to 262,144 tokens. At Q4_K_M it needs about 3.14 GB of VRAM — see which GPUs and Macs can run it below.
1.9K downloads
59 likesSpecifications
- Publisher
- janhq
- Parameters
- 4.4B
- Architecture
- Qwen3ForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 151,936
- Release Date
- 2026-01-19
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Jan v3 4B Base Instruct Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 2.4 GB | 26.3 GB | 1.87 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 2.4 GB | 26.4 GB | 1.93 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 2.6 GB | 26.6 GB | 2.15 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 2.7 GB | 26.7 GB | 2.21 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 3.1 GB | 27.1 GB | 2.65 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 3.6 GB | 27.6 GB | 3.14 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 4.1 GB | 28.1 GB | 3.64 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 4.9 GB | 28.9 GB | 4.41 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Jan v3 4B Base Instruct?
Q4_K_M · 3.1 GBJan v3 4B Base Instruct (Q4_K_M) requires 3.1 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 5+ GB is recommended. Using the full 262K context window can add up to 24.0 GB, bringing total usage to 27.1 GB. 50 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run Jan v3 4B Base Instruct?
Q4_K_M · 3.1 GB59 devices with unified memory can run Jan v3 4B Base Instruct, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomWhere to Download Jan v3 4B Base Instruct
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Related Models
Frequently Asked Questions
- How much VRAM does Jan v3 4B Base Instruct need?
Jan v3 4B Base Instruct requires 3.1 GB of VRAM at Q4_K_M, or 9.3 GB at BF16. Full 262K context adds up to 24.0 GB (27.1 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 4.4B × 4.8 bits ÷ 8 = 2.6 GB
KV Cache + Overhead ≈ 0.5 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 24.5 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M3.1 GBQ4_K_M + full context27.1 GB- What's the best quantization for Jan v3 4B Base Instruct?
For Jan v3 4B Base Instruct, Q4_K_M (3.1 GB) offers the best balance of quality and VRAM usage. Q4_K_L (3.2 GB) provides better quality if you have the VRAM. The smallest option is IQ2_M at 2.0 GB.
VRAM requirement by quantization
IQ2_M2.0 GBQ3_K_M2.6 GBQ4_K_S3.0 GBQ4_K_M ★3.1 GBQ5_K_S3.5 GBBF169.3 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Jan v3 4B Base Instruct on a Mac?
Jan v3 4B Base Instruct requires at least 2.0 GB at IQ2_M, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Jan v3 4B Base Instruct locally?
Yes — Jan v3 4B Base Instruct can run locally on consumer hardware. At Q4_K_M quantization it needs 3.1 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Jan v3 4B Base Instruct?
At Q4_K_M, Jan v3 4B Base Instruct can reach ~1401 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~209 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 ÷ 3.1 × 0.65 = ~1656 tok/s
Estimated speed at Q4_K_M (3.1 GB)
~1656 tok/s~209 tok/s~1656 tok/s~1401 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Jan v3 4B Base Instruct?
At Q4_K_M, the download is about 2.65 GB. The full-precision BF16 version is 8.82 GB. The smallest option (IQ2_M) is 1.49 GB.
- Which GPUs can run Jan v3 4B Base Instruct?
50 consumer GPUs can run Jan v3 4B Base Instruct at Q4_K_M (3.1 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 50 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Jan v3 4B Base Instruct?
59 devices with unified memory can run Jan v3 4B Base Instruct at Q4_K_M (3.1 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Apple iPhone 17 Pro, Asus ROG Flow Z13 (2025, 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.