Phi-tiny-MoE-instruct

Home/Models/microsoft/Phi-tiny-MoE-instruct

microsoft/Phi-tiny-MoE-instruct

kind: original
quant: bf16 (16-bit)
family: misc
modality: text-generation
root model: microsoft/Phi-tiny-MoE-instruct
provider: microsoft
license: mit
language: en
context: 4k
library: transformers
vRAM (GB): —
size on disk: —
downloads: 880,571
likes: 39
created: —
last modified: 2025-12-10T18:21:11+00:00 (2025)
huggingface: https://huggingface.co/microsoft/Phi-tiny-MoE-instruct
tags: transformerssafetensorsphimoetext-generationconversationalcustom_codeenarxiv:2506.18349arxiv:2404.14219arxiv:2409.12136license:mitregion:us

benchmark	category	score	conf	evidence
GSM8K	math	78.47	high	`\| Phi-tiny-MoE \| 3.8B \| 1.1B \| 60.83 \| 36.34 \| 45.58 \| 76.37 \| 58.50 \| 78.47 \| 7.05 \|`
MMLU	knowledge	60.83	high	`\| Phi-tiny-MoE \| 3.8B \| 1.1B \| 60.83 \| 36.34 \| 45.58 \| 76.37 \| 58.50 \| 78.47 \| 7.05 \|`
BBH	reasoning	45.58	high	`\| Phi-tiny-MoE \| 3.8B \| 1.1B \| 60.83 \| 36.34 \| 45.58 \| 76.37 \| 58.50 \| 78.47 \| 7.05 \|`
MMLU-Pro	knowledge	36.34	high	`\| Phi-tiny-MoE \| 3.8B \| 1.1B \| 60.83 \| 36.34 \| 45.58 \| 76.37 \| 58.50 \| 78.47 \| 7.05 \|`
MT-Bench	chat	7.05	high	`\| Phi-tiny-MoE \| 3.8B \| 1.1B \| 60.83 \| 36.34 \| 45.58 \| 76.37 \| 58.50 \| 78.47 \| 7.05 \|`

none scraped.