Skip to content
← Models
Open weightsDatabricks/10 trustMetrics-checked

Dolly 2.0

April 2023's first truly open, commercially usable instruction-tuned LLM.

Visit official page →

About

Dolly 2.0, released by Databricks on April 12, 2023, was a milestone in open AI: a 12B instruction-following model built on EleutherAI's Pythia-12B and fine-tuned on databricks-dolly-15k — 15,000 instruction-response pairs written from scratch by roughly 5,000 Databricks employees, released under CC-BY-SA.

What made it historic was its clean IP story: unlike Alpaca and other early instruction models that distilled OpenAI outputs (and inherited their terms-of-service restrictions), every component — weights, code, and data — was cleared for commercial use. Databricks framed it as proof that organizations could build ChatGPT-style interactivity without API dependence or data sharing.

It is a legacy model now, far behind modern instruction-tuned models in capability, but an important reference point in open-model licensing history.

Specs

Context window

Pricing

Free quota

API access

Weights

Open weights

Released

April 12, 2023

Benchmarks

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

Use cases

instruction-tuningopen-licensingai-history

Trust score — 6/10

/10 trustMetrics-checked

Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 76; Trending score 58/100 from researcher assessment.

More in the dolly-v2 family

Sources