utM-A-P · Downloadable
OpenCodeInterpreter-SC2-3B
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Feb 29, 2024
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
YOUR MEMORY BUDGET
Local AI models for 8GB VRAM.
Explore models by theoretical weight storage. This shortlist covers available downloadable language models with a known or name-derived parameter count.
6 GBleft for weights at 4-bit
8 GB budget − 2 GB reserve
8 GB budget − 2 GB reserve
The reserve is your assumption, not a measured overhead allowance. The default 2 GB may be too little for your runtime or context length. Results are not confirmed fits and have no verified speed rating. Weight estimates use decimal GB and do not include quantization overhead. For timed results on one 12GB machine, see measured speeds.
852 models within the weight budget
Largest parameter count first · not a quality rankingMeta · Downloadable
Llama-3.2-3B
Generates text for conversations and writing.
- Source listed
- Sep 18, 2024
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- llama3.2
Meta · Downloadable
Llama-3.2-3B-Instruct
Generates text for conversations and writing.
- Source listed
- Sep 18, 2024
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- llama3.2
Microsoft · Downloadable
Dayhoff-3b-GR-HM
Generates text for conversations and writing.
- Source listed
- Jul 4, 2025
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- mit
Microsoft · Downloadable
Dayhoff-3b-GR-HM-c
Generates text for conversations and writing.
- Source listed
- Jul 4, 2025
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- mit
Microsoft · Downloadable
Dayhoff-3b-UR90
Generates text for conversations and writing.
- Source listed
- Jul 4, 2025
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- mit
Microsoft · Downloadable
Dayhoff-3b-UR90-10
Generates text for conversations and writing.
- Source listed
- Apr 14, 2026
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- mit
Microsoft · Downloadable
GUI-Actor-3B-Qwen2.5-VL
Understands text and images and responds in text.
- Source listed
- Jun 1, 2025
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- mit
OpenBMB · Downloadable
BitCPM-CANN-3B
Generates text for conversations and writing.
- Source listed
- May 15, 2026
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
OpenBMB · Downloadable
NOSA-3B
Generates text for conversations and writing.
- Source listed
- Jan 6, 2026
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
Stability AI · Downloadable
japanese-stablelm-3b-4e1t-base
Generates text for conversations and writing.
- Source listed
- Oct 16, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
Stability AI · Downloadable
japanese-stablelm-3b-4e1t-instruct
Generates text for conversations and writing.
- Source listed
- Oct 16, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
Stability AI · Downloadable
stable-code-3b
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Jan 9, 2024
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- other
Stability AI · Downloadable
stable-code-instruct-3b
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Mar 6, 2024
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- other
Stability AI · Downloadable
stablecode-completion-alpha-3b
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Jul 31, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
Stability AI · Downloadable
stablecode-completion-alpha-3b-4k
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Aug 7, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- apache-2.0
Stability AI · Downloadable
stablecode-instruct-alpha-3b
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Aug 7, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- other
Stability AI · Downloadable
stablelm-3b-4e1t
Generates text for conversations and writing.
- Source listed
- Sep 29, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- cc-by-sa-4.0
Stability AI · Downloadable
stablelm-base-alpha-3b
Generates text for conversations and writing.
- Source listed
- Apr 17, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- cc-by-sa-4.0
Stability AI · Downloadable
stablelm-base-alpha-3b-v2
Generates text for conversations and writing.
- Source listed
- Aug 4, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- cc-by-sa-4.0
Stability AI · Downloadable
stablelm-tuned-alpha-3b
Generates text for conversations and writing.
- Source listed
- Apr 19, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- cc-by-nc-sa-4.0
Stability AI · Downloadable
stablelm-zephyr-3b
Generates text for conversations and writing.
- Source listed
- Nov 21, 2023
- Parameters
- 3B
- 4-bit weights only
- ≈ 1.5 GB
- License
- other
Meta · Downloadable
MobileMoE-M-Base
Generates text for conversations and writing.
- Source listed
- Jun 11, 2026
- Parameters
- 2.8B
- 4-bit weights only
- ≈ 1.4 GB
- License
- fair-noncommercial-research-license
Meta · Downloadable
MobileMoE-M-SFT
Generates text for conversations and writing.
- Source listed
- Jun 11, 2026
- Parameters
- 2.8B
- 4-bit weights only
- ≈ 1.4 GB
- License
- fair-noncommercial-research-license
What changes with a 8GB budget?
With 2 GB set aside, 6 GB remains for weights. At 4-bit that allows up to 12 billion parameters in the theoretical calculation. An 8-billion-parameter model alone needs approximately 4 GB at this precision, before its conversation cache and runtime.
Why might a listed model still fail to load?
Its actual weight format may need more space, the runtime may not support its architecture, or the conversation cache may exceed your reserve. Vision models can also need space for image processing. Check the original model card and measure your exact setup before relying on it.
Read the memory guide and technical sources · Compare two models