Autotrain compatible Endpoints compatible Merge Mixtral Moe Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/MaziyarPanahi/Topxtral-4x7B-v0.1
Topxtral 4x7B V0.1 Benchmarks
ARC: 72.53 vs 96.7 (so35)-25%
HellaSwag: 88.33 vs 95.3 (gpt4)-7.3%
MMLU: 64.96 vs 88.3 (so35)-26.4%
TruthfulQA: 73.38 vs 59 (gpt4)24.4%
WinoGrande: 83.19 vs 87.5 (gpt4)-4.9%
GSM8K: 71.72 vs 96.4 (so35)-25.6%
LLME Score: 0.19218
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)
Topxtral 4x7B V0.1 Parameters and Internals
| Additional Notes |
|
| LLM Name | Topxtral 4x7B V0.1 |
| Repository 🤗 | https://huggingface.co/MaziyarPanahi/Topxtral-4x7B-v0.1 |
| Model Size | 18.5b |
| Required VRAM | 37.1 GB |
| Updated | 2025-07-30 |
| Maintainer | MaziyarPanahi |
| Model Type | mixtral |
| Model Files | |
| Model Architecture | MixtralForCausalLM |
| License | apache-2.0 |
| Context Length | 32768 |
| Model Max Length | 32768 |
| Transformers Version | 4.37.2 |
| Tokenizer Class | LlamaTokenizer |
| Padding Token | <s> |
| Vocabulary Size | 32000 |
| Torch Data Type | bfloat16 |
Best Alternatives to Topxtral 4x7B V0.1
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| EastAsia 4x7B MoE Experiment | 32K / 37.1 GB | 1208 | 1 |
| Lumina 3.5 | 32K / 37.1 GB | 2118 | 0 |
| Hyperion 3.0 Mixtral 3x7B | 32K / 37.1 GB | 4 | 4 |
| Blitz AI MoE V0.7 | 32K / 37.1 GB | 4 | 1 |
| Blitz AI MoE V0.4 | 32K / 37.1 GB | 5 | 1 |
| HeroBophades 3x7B | 32K / 37.1 GB | 10 | 3 |
| Wizard Kun Lake 3x7B MoE | 32K / 37.1 GB | 7 | 1 |
| NaruMOE 3x7B V2 | 32K / 37.1 GB | 9 | 0 |
| MoE 3x7b QA Code Inst | 32K / 37 GB | 8 | 4 |
| Pearl 3x7B | 32K / 37.1 GB | 6 | 2 |
Rank the Topxtral 4x7B V0.1 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? 50262 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124