utDeepSeek · Downloadable
deepseek-coder-33b-base
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Oct 28, 2023
- Parameters
- 33B
- 4-bit weights only
- ≈ 16.5 GB
- License
- other
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 rankingDeepSeek · Downloadable
deepseek-coder-33b-instruct
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Nov 1, 2023
- Parameters
- 33B
- 4-bit weights only
- ≈ 16.5 GB
- License
- other
M-A-P · Downloadable
OpenCodeInterpreter-DS-33B
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- Feb 19, 2024
- Parameters
- 33B
- 4-bit weights only
- ≈ 16.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
GroveMoE-Base
Generates text for conversations and writing.
- Source listed
- Aug 18, 2025
- Parameters
- 32.9B
- 4-bit weights only
- ≈ 16.5 GB
- License
- apache-2.0
inclusionAI · Downloadable
GroveMoE-Inst
Generates text for conversations and writing.
- Source listed
- Aug 18, 2025
- Parameters
- 32.9B
- 4-bit weights only
- ≈ 16.5 GB
- License
- apache-2.0
Tencent · Downloadable
DRIVE-RL
Generates text for conversations and writing.
- Source listed
- Nov 12, 2025
- Parameters
- 32.8B
- 4-bit weights only
- ≈ 16.4 GB
- License
- Not disclosed
Tencent · Downloadable
DRIVE-SFT
Generates text for conversations and writing.
- Source listed
- Nov 12, 2025
- Parameters
- 32.8B
- 4-bit weights only
- ≈ 16.4 GB
- License
- Not disclosed
Meta · Downloadable
cwm
Generates text, according to its published model architecture (CwmForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Aug 25, 2025
- Parameters
- 32.6B
- 4-bit weights only
- ≈ 16.3 GB
- License
- fair-noncommercial-research-license
Meta · Downloadable
cwm-sft
Generates text, according to its published model architecture (CwmForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Aug 25, 2025
- Parameters
- 32.6B
- 4-bit weights only
- ≈ 16.3 GB
- License
- fair-noncommercial-research-license
DeepSeek · Downloadable
DeepSeek-R1-Distill-Qwen-32B
Generates text for conversations and writing. Supports a reasoning mode.
- Source listed
- Jan 20, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- mit
inclusionAI · Downloadable
AReaL-boba-2-32B
Generates text for conversations and writing.
- Source listed
- Jun 3, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
inclusionAI · Downloadable
AReaL-boba-SFT-32B
Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Mar 29, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
inclusionAI · Downloadable
Qwen3-32B-AWorld
Generates text for conversations and writing.
- Source listed
- Aug 29, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
M-A-P · Downloadable
CriticLeanGPT-Qwen2.5-32B-Instruct-SFT
Generates text for conversations and writing.
- Source listed
- Jul 11, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- other
M-A-P · Downloadable
CriticLeanGPT-Qwen2.5-32B-Instruct-SFT-RL
Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Jul 11, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- Not disclosed
M-A-P · Downloadable
CriticLeanGPT-Qwen2.5-32B-RL
Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Jul 12, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- Not disclosed
M-A-P · Downloadable
CriticLeanGPT-Qwen3-32B-RL
Generates text, according to its published model architecture (Qwen3ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Jul 10, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- Not disclosed
M-A-P · Downloadable
OProver-32B
Generates text for conversations and writing.
- Source listed
- May 15, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
M-A-P · Downloadable
OProver-32B-Base
Generates text for conversations and writing.
- Source listed
- May 15, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
M-A-P · Downloadable
OProver-32B-Round1
Generates text for conversations and writing.
- Source listed
- May 15, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
M-A-P · Downloadable
TerminalTraj-32B
Generates text, according to its published model architecture (Qwen2ForCausalLM). No task is listed on Hugging Face, so check the model card for intended use.
- Source listed
- Feb 10, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- Not disclosed
Microsoft · Downloadable
AesCode-32B
Designed for coding and software tasks. Understands text and images and responds in text.
- Source listed
- Sep 29, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- apache-2.0
Microsoft · Downloadable
FrogBoss-32B-2510
Generates text for conversations and writing.
- Source listed
- Jan 5, 2026
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- mit
Microsoft · Downloadable
NextCoder-32B
Designed for coding and software tasks. Generates text for conversations and writing.
- Source listed
- May 3, 2025
- Parameters
- 32B
- 4-bit weights only
- ≈ 16 GB
- License
- mit
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