LLM EXPLORER
 LLM Explorer: A Curated Large Language Model Directory and Analytics
Dark Theme
  LLM  List    LLM Hosting    LLM Leaderboards    Blog    Newsfeed    Advertise

TinyMistral 248M V2.5 by Locutusque

 »  All LLMs  »  Locutusque  »  TinyMistral 248M V2.5   URL Share it on

  5   
  Merged Model   Autotrain compatible   Code   Computer science Dataset:open-phi/programming b...   Dataset:open-phi/textbooks   En   Endpoints compatible   Mistral   Model-index   Region:us   Safetensors
Model Card on HF 🤗: https://huggingface.co/Locutusque/TinyMistral-248M-v2.5 

TinyMistral 248M V2.5 Benchmarks

MMLU Pro: 1.5
GPQA: 0.11
MUSR: 5.07
BBH: 3.18
IFEval: 13.36 vs 88 (so35)-84.8%
ARC: 24.57 vs 96.7 (so35)-74.6%
HellaSwag: 27.49 vs 95.3 (gpt4)-71.2%
MMLU: 23.15 vs 88.3 (so35)-73.8%
TruthfulQA: 46.72 vs 59 (gpt4)-20.8%
WinoGrande: 47.83 vs 87.5 (gpt4)-45.3%
MATH Lvl 5: 0.98
LLME Score: 0.19405
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
TinyMistral 248M V2.5 (Locutusque/TinyMistral-248M-v2.5)
🌟 Advertise your project 🚀

TinyMistral 248M V2.5 Parameters and Internals

Model Type 
text-generation
Additional Notes 
This model fixes exploding gradient issues present in v2, making it easier to fine-tune.
Training Details 
Data Sources:
open-phi/programming_books_llama, open-phi/textbooks
Methodology:
The model was created by merging TinyMistral-248M-v1 and v2, then further pretraining on synthetic textbooks.
Input Output 
Performance Tips:
The model's performance seems to degrade in the inference API. It's recommended to try the model by installing it directly.
LLM NameTinyMistral 248M V2.5
Repository 🤗https://huggingface.co/Locutusque/TinyMistral-248M-v2.5 
Merged ModelYes
Model Size248m
Required VRAM1 GB
Updated2025-10-13
MaintainerLocutusque
Model Typemistral
Model Files  1.0 GB
Supported Languagesen code
Model ArchitectureMistralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.36.2
Tokenizer ClassLlamaTokenizer
Padding Token<|endoftext|>
Vocabulary Size32005
Torch Data Typefloat16

Best Alternatives to TinyMistral 248M V2.5

Best Alternatives
Context / RAM
Downloads
Likes
TinyMistral 248M V232K / 1 GB83117
TinyMistral 248M Instruct32K / 1 GB138012
...adrin TinyMistral248M Instruct32K / 0.5 GB7767
TinyMistral 248M V2.5 Instruct32K / 1 GB211
...al V2.5 MiniPile Guidelines E132K / 0.6 GB72
TinyMistral V2 Test132K / 1 GB181
TinyMistral 248M 8bits32K / 0.3 GB441
Tinymistv132K / 0.5 GB170
TinyMistral Haiku32K / 1 GB50
...istral 248M V2.5 Instruct Orpo32K / 0.5 GB70

Rank the TinyMistral 248M V2.5 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! 🌟

Instruction Following and Task Automation  
Factuality and Completeness of Knowledge  
Censorship and Alignment  
Data Analysis and Insight Generation  
Text Generation  
Text Summarization and Feature Extraction  
Code Generation  
Multi-Language Support and Translation  

What open-source LLMs or SLMs are you in search of? 51544 in total.

Email us: info@extractum.io. Our Privacy Policy  | Terms and Conditions  | Suggest an improvement.
Our Social Media →  
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124
  Support LLM Explorer