MDL
Phi-3 Mini Instruct 3.8B
Microsoft
Verified against Artificial Analysis · Jul 13, 2026
Capability
24
percentile index
Price /M
$0 / $0
in / out
Context
—
Avg score
15.7
11 benchmarks
AA Intelligence
4.6
index
Weights
Open
Released
Apr 23, 2024
About
Phi-3 Mini Instruct 3.8B is a open-weights model in the Phi family from Microsoft. Benchmarked on 11 evals, averaging 15.7.
Benchmark scores · 11
avg 15.7 — higher is betterThe research behind this model
all papers mentioning it →Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone151 cites · Apr 22, 2024BioRefusalAudit: Auditing Biosecurity Refusal Depth Using General and Domain-Fine-Tuned Sparse AutoencodersMay 28, 2026SlimMoE: Structured Compression of Large MoE Models via Expert Slimming and DistillationJun 23, 2025
Mentions matched by name in title/abstract — from the arXiv + HF daily corpus.
History
full ledger →Capability over time
Append-only ledger — every observed change to this model's numbers.
API providers
Artificial Analysis$0/M