Arxiv:2401.15006 Autotrain compatible Dataset:ai4bharat/indic-instru... En Endpoints compatible Gguf Hi Instruct Llama Llama2 Model-index Multilingual Quantized Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/ai4bharat/Airavata
Airavata Benchmarks
MMLU Pro: 7.05
GPQA: 3.24
MUSR: 4.04
BBH: 11.57
IFEval: 5.59 vs 88 (so35)-93.6%
ARC: 46.5 vs 96.7 (so35)-51.9%
HellaSwag: 69.26 vs 95.3 (gpt4)-27.3%
MMLU: 43.9 vs 88.3 (so35)-50.3%
TruthfulQA: 40.62 vs 59 (gpt4)-31.2%
WinoGrande: 68.82 vs 87.5 (gpt4)-21.3%
GSM8K: 4.02 vs 96.4 (so35)-95.8%
MATH Lvl 5: 1.81
LLME Score: 0.20503
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)
Airavata Parameters and Internals
| Model Type |
| |||||||||
| Use Cases |
| |||||||||
| Supported Languages |
| |||||||||
| Training Details |
| |||||||||
| Input Output |
|
| LLM Name | Airavata |
| Repository 🤗 | https://huggingface.co/ai4bharat/Airavata |
| Model Size | 6.9b |
| Required VRAM | 13.8 GB |
| Updated | 2025-08-18 |
| Maintainer | ai4bharat |
| Model Type | llama |
| Instruction-Based | Yes |
| Model Files | |
| Supported Languages | en hi |
| GGUF Quantization | Yes |
| Quantization Type | gguf |
| Model Architecture | LlamaForCausalLM |
| License | llama2 |
| Context Length | 4096 |
| Model Max Length | 4096 |
| Transformers Version | 4.37.1 |
| Tokenizer Class | LlamaTokenizer |
| Padding Token | <pad> |
| Vocabulary Size | 48065 |
| Torch Data Type | bfloat16 |
Rank the Airavata 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? 50729 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124