Merged Model Autotrain compatible Base model:finetune:microsoft/... Base model:microsoft/phi-4 Conversational Custom code Endpoints compatible Phi3 Region:us Safetensors Sharded Tensorflow
Model Card on HF 🤗: https://huggingface.co/ehristoforu/phi-4-25b
Phi 4 25B Benchmarks
MMLU Pro: 48.34
GPQA: 9.17
MUSR: 11.43
BBH: 55.67
IFEval: 64.84 vs 88 (so35)-26.3%
MATH Lvl 5: 45.24
LLME Score: 0.31469
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)
Phi 4 25B Parameters and Internals
| LLM Name | Phi 4 25B |
| Repository 🤗 | https://huggingface.co/ehristoforu/phi-4-25b |
| Base Model(s) | |
| Merged Model | Yes |
| Model Size | 25b |
| Required VRAM | 49.8 GB |
| Updated | 2025-09-23 |
| Maintainer | ehristoforu |
| Model Type | phi3 |
| Model Files | |
| Model Architecture | Phi3ForCausalLM |
| Context Length | 16384 |
| Model Max Length | 16384 |
| Transformers Version | 4.46.2 |
| Tokenizer Class | GPT2Tokenizer |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 100352 |
| Torch Data Type | bfloat16 |
Best Alternatives to Phi 4 25B
Best Alternatives |
Context / RAM |
Downloads |
Likes |
|---|---|---|---|
| Sasando 1 25M | 0K / 0.1 GB | 60 | 1 |
Rank the Phi 4 25B 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? 51544 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124