Outlier 70B V3.2 is an open-source language model by Outlier-Ai. Features: 32b LLM, Context: 32K, License: apache-2.0, Instruction-Based, LLM Explorer Score: 0.32.
1.58-bit Base model:finetune:qwen/qwen2... Base model:qwen/qwen2.5-32b-in... Bitnet-alternative Consumer-hardware Conversational Custom code Efficient-inference En Instruct Low-memory Mixture-of-experts Moe Outlier moe Qwen2.5 Region:us Superseded Ternary Ternary-quantized
Model Card on HF 🤗: https://huggingface.co/Outlier-Ai/Outlier-70B-V3.2
Outlier 70B V3.2 Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Outlier 70B V3.2 Parameters and Internals
| LLM Name | Outlier 70B V3.2 |
| Repository 🤗 | https://huggingface.co/Outlier-Ai/Outlier-70B-V3.2 |
| Base Model(s) | |
| Model Size | 32b |
| Required VRAM | 0 GB |
| Updated | 2026-04-14 |
| Maintainer | Outlier-Ai |
| Model Type | outlier_moe |
| Instruction-Based | Yes |
| Model Files | |
| Supported Languages | en |
| Model Architecture | OutlierMoEForCausalLM |
| License | apache-2.0 |
| Context Length | 32768 |
| Model Max Length | 32768 |
| Transformers Version | 4.43.1 |
| Tokenizer Class | Qwen2Tokenizer |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 152064 |
| Torch Data Type | bfloat16 |
| Errors | replace |
Rank the Outlier 70B V3.2 Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 52861 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a