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

Saka 7.6B by Sakalti

 »  All LLMs  »  Sakalti  »  Saka 7.6B   URL Share it on

  11   

Saka 7.6B is an open-source language model by Sakalti. Features: 7b LLM, VRAM: 16.2GB, Context: 32K, Merged, LLM Explorer Score: 0.26.

  Merged Model   Arxiv:2306.01708 Base model:sousiomine/kuroiso-... Base model:suayptalha/homercre...   Conversational   Endpoints compatible   Qwen2   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Sakalti/Saka-7.6B 

Saka 7.6B Benchmarks

MMLU Pro: 39.34
GPQA: 8.84
MUSR: 14.95
BBH: 37.97
IFEval: 45.24 vs 88 (so35)-48.6%
MATH Lvl 5: 32.55
LLME Score: 0.26444
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
Saka 7.6B (Sakalti/Saka-7.6B)
🌟 Advertise your project 🚀

Saka 7.6B Parameters and Internals

LLM NameSaka 7.6B
Repository 🤗https://huggingface.co/Sakalti/Saka-7.6B 
Base Model(s)  HomerCreativeAnvita Mix Qw7B   Kuroiso CR 7B 20250124   suayptalha/HomerCreativeAnvita-Mix-Qw7B   SousiOmine/Kuroiso-CR-7B-20250124
Merged ModelYes
Model Size7b
Required VRAM16.2 GB
Updated2026-04-10
MaintainerSakalti
Model Typeqwen2
Model Files  1.1 GB: 1-of-30   1.1 GB: 2-of-30   0.5 GB: 3-of-30   0.5 GB: 4-of-30   0.5 GB: 5-of-30   0.5 GB: 6-of-30   0.5 GB: 7-of-30   0.5 GB: 8-of-30   0.5 GB: 9-of-30   0.5 GB: 10-of-30   0.5 GB: 11-of-30   0.5 GB: 12-of-30   0.5 GB: 13-of-30   0.5 GB: 14-of-30   0.5 GB: 15-of-30   0.5 GB: 16-of-30   0.5 GB: 17-of-30   0.5 GB: 18-of-30   0.5 GB: 19-of-30   0.5 GB: 20-of-30   0.5 GB: 21-of-30   0.5 GB: 22-of-30   0.5 GB: 23-of-30   0.5 GB: 24-of-30   0.5 GB: 25-of-30   0.5 GB: 26-of-30   0.5 GB: 27-of-30   0.5 GB: 28-of-30   0.5 GB: 29-of-30   0.5 GB: 30-of-30
Model ArchitectureQwen2ForCausalLM
Context Length32768
Model Max Length32768
Transformers Version4.46.3
Tokenizer ClassQwen2Tokenizer
Padding Token<|PAD_TOKEN|>
Vocabulary Size152064
Torch Data Typefloat16
Errorsreplace

Best Alternatives to Saka 7.6B

Best Alternatives
Context / RAM
Downloads
Likes
Qwen2.5 7B Preview986K / 15.2 GB50
Qwen2.5 7B Instruct 1M986K / 15.4 GB101838363
Hush Qwen2.5 7B V1.2986K / 15.2 GB31
Hush Qwen2.5 7B V1.1986K / 15.2 GB61
Hush Qwen2.5 7B V1.4986K / 15.2 GB41
Hush Qwen2.5 7B Preview986K / 15.2 GB50
Hush Qwen2.5 7B V1.3986K / 15.2 GB42
Hush Qwen2.5 7B RP V1.4 1M986K / 15.2 GB42
Qwen 2.5 7B Exp Sce986K / 15.2 GB82
Qwen2.5 7B MixStock V0.1986K / 15.2 GB303
Note: green Score (e.g. "73.2") means that the model is better than Sakalti/Saka-7.6B.

Rank the Saka 7.6B 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? 52721 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.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a
  Support LLM Explorer