Qwen/Qwen2.5-1.5B-Instruct

In plain English

Good for a quick first read of US style contracts and flagging clauses worth a closer look. Has no specialist US legal training, so it is not a substitute for a licensed attorney and should not be used for filings.

A 1.5 billion parameter instruction tuned model, light enough for quick local experimentation while still handling multi step instructions reasonably well.

Required documentation

Complete

Intended use

Intended for a first pass read of general contracts to flag clauses worth a closer look. Not intended to replace review by a licensed attorney.

Training data provenance

Qwen's instruction-tuned checkpoint trained on a broad web and multilingual corpus, with no specific US legal training data.

Licence & terms

Apache-2.0.

Reviews3

Write a review

MRMaya Rehanirep 4.8Research989aa79

7h ago

Accuracy3.2

Fine-tune3.3

Documentation3.4

Speed4.2

TNTomasz Nowakrep 4.5Prototype989aa79

7h ago

Accuracy3.0

Fine-tune3.1

Documentation3.2

Speed4.4

PSPriya Shahrep 4.6Research989aa79

7h ago

Accuracy3.3

Fine-tune3.4

Documentation3.5

Speed4.3

Discussion0

Posting as a signed-in userSign in to post

No questions yet. Be the first to ask one.

Structured rating

3.5

from 3 reviews of 989aa79

Only 3 reviews on 989aa79 so far. Earlier versions carry more, but they are not this build.

Accuracy3.2

Fine-tune3.3

Documentation3.4

Speed4.3

Run it now

ollama run Qwen/Qwen2.5-1.5B-Instruct

Will it fit?

Build	File	VRAM
Q4_K_M	986 MB	2.0 GB
Q8_0	1.6 GB	2.6 GB

Set a hardware profile in the browse sidebar to see which builds fit.

Inference providers

No hosted providers listed yet.

No cloud required. Run it with

vLLMSGLang

Model tree

Base modelQwen/Qwen2.5-1.5B

Adapters0

Finetunes0

Quantizations0

Merges0

At a glance

Parameters—

Licenceapache-2.0

Downloads7.3M

Reviews3

Docs completenessComplete