utMicrosoft · Downloadable
UniRG-CXR
Understands text and images and responds in text.
- Source listed
- Mar 19, 2026
- Parameters
- 8.8B
- 4-bit weights only
- ≈ 4.4 GB
- License
- apache-2.0
YOUR MEMORY BUDGET
Local AI models for 24GB VRAM.
Explore models by theoretical weight storage. This shortlist covers available downloadable language models with a known or name-derived parameter count.
22 GBleft for weights at 4-bit
24 GB budget − 2 GB reserve
24 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.
1,129 models within the weight budget
Largest parameter count first · not a quality rankingNVIDIA · Downloadable
nemotron-labs-audio-visual-flamingo-hf
Generates text for conversations and writing.
- Source listed
- Apr 15, 2026
- Parameters
- 8.7B
- 4-bit weights only
- ≈ 4.4 GB
- License
- other
OpenBMB · Downloadable
MiniCPM-V-4_5
Understands text and images and responds in text.
- Source listed
- Aug 24, 2025
- Parameters
- 8.7B
- 4-bit weights only
- ≈ 4.3 GB
- License
- apache-2.0
OpenBMB · Downloadable
MiniCPM-Llama3-V-2_5
Understands text and images and responds in text.
- Source listed
- May 19, 2024
- Parameters
- 8.5B
- 4-bit weights only
- ≈ 4.3 GB
- License
- Not disclosed
BAAI · Downloadable
Emu3-Chat
Generates text for conversations and writing.
- Source listed
- Sep 26, 2024
- Parameters
- 8.5B
- 4-bit weights only
- ≈ 4.2 GB
- License
- apache-2.0
inclusionAI · Downloadable
ViLaSR
Understands text and images and responds in text.
- Source listed
- Jun 1, 2025
- Parameters
- 8.3B
- 4-bit weights only
- ≈ 4.1 GB
- License
- Not disclosed
BAAI · Downloadable
AquilaMed-RL
Generates text for conversations and writing.
- Source listed
- Jun 12, 2024
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- other
inclusionAI · Downloadable
Ling-3.0-tiny-base
Generates text, according to its published model architecture (BailingMoeV3ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Aug 11, 2026
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- mit
OpenBMB · Downloadable
SciCore-Mol
Generates text for conversations and writing.
- Source listed
- Apr 29, 2026
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
AgentCPM-Report
Generates text for conversations and writing.
- Source listed
- Jan 19, 2026
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
InfLLM-V2-Long-Sparse-Base
Generates text, according to its published model architecture (MiniCPMForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Oct 24, 2025
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
InfLLM-V2-Short-Dense-Base
Generates text, according to its published model architecture (MiniCPMForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Oct 25, 2025
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
MiniCPM4-MCP
Generates text for conversations and writing.
- Source listed
- Jun 5, 2025
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
MiniCPM4-Survey
Generates text for conversations and writing.
- Source listed
- Jun 5, 2025
- Parameters
- 8.2B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
BAAI · Downloadable
Video-XL-2
Generates text for conversations and writing.
- Source listed
- May 30, 2025
- Parameters
- 8.1B
- 4-bit weights only
- ≈ 4.1 GB
- License
- apache-2.0
OpenBMB · Downloadable
AgentCPM-GUI
Understands text and images and responds in text.
- Source listed
- May 8, 2025
- Parameters
- 8.1B
- 4-bit weights only
- ≈ 4 GB
- License
- apache-2.0
OpenBMB · Downloadable
MiniCPM-V-2_6
Understands text and images and responds in text.
- Source listed
- Aug 4, 2024
- Parameters
- 8.1B
- 4-bit weights only
- ≈ 4 GB
- License
- Not disclosed
Meta · Downloadable
meta-wiki-expert
Generates text, according to its published model architecture (LlamaForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Aug 5, 2025
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- cc-by-nc-4.0
Aleph Alpha · Downloadable
llama-3_1-8b-tfree-hat-base
Generates text, according to its published model architecture (HATForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Apr 25, 2025
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- other
Aleph Alpha · Downloadable
llama-3_1-8b-tfree-hat-dpo
Generates text, according to its published model architecture (HATForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Apr 25, 2025
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- other
Aleph Alpha · Downloadable
llama-3_1-8b-tfree-hat-sft
Generates text, according to its published model architecture (HATForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Apr 25, 2025
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- other
BAAI · Downloadable
Bunny-Llama-3-8B-V
Generates text for conversations and writing.
- Source listed
- Apr 21, 2024
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- apache-2.0
BAAI · Downloadable
Bunny-v1_1-Llama-3-8B-V
Understands text and images and responds in text.
- Source listed
- May 31, 2024
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- apache-2.0
BAAI · Downloadable
Infinity-Instruct-3M-0625-Llama3-8B
Generates text for conversations and writing.
- Source listed
- Jul 9, 2024
- Parameters
- 8B
- 4-bit weights only
- ≈ 4 GB
- License
- apache-2.0
What changes with a 24GB budget?
With 2 GB set aside, 22 GB remains for weights. At 4-bit that allows up to 44 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