Skip to content
← Models
Open weightsMicrosoft/10 trustVerified

Phi-1.5

Microsoft's 1.3B textbook-trained small language model.

Visit official page →

About

Phi-1.5 is a 1.3-billion-parameter Transformer from Microsoft Research, released in November 2023 as the follow-up to Phi-1. It was trained on the same synthetic textbook-quality sources as Phi-1 plus new NLP synthetic texts, giving it common-sense reasoning and language understanding comparable to models five times larger.

It was released as a base model without instruction tuning or RLHF, under the MIT license, as a research playground for studying safety, toxicity, and fine-tuning in small models. It kicked off the Phi small-language-model line that continued through Phi-2, Phi-3, and Phi-4.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

November 1, 2023

Benchmarks

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

researchon-device-aireasoning

Trust score — 8/10

/10 trustVerified

Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 126; Trending score 78/100 from researcher assessment.

More in the phi family

Sources