NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #5620 most downloaded on PyPI
DSPy
Last release 10 days ago
25 Sep 2026
Release timing varies
gaps range from 2 weeks to 2 months
Nearly every release is documented
notes for 60 of the last 60 stable releases
7 versions withdrawn
withdrawn after publishing
3 years old
143 releases · first in 2023
One column per quarter.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Various improvements for adapters & LMs by @okhat in https://github.com/stanfordnlp/dspy/pull/1523
Full Changelog: https://github.com/stanfordnlp/dspy/compare/2.4.16...2.4.17
This is a pretty significant release. New dspy.LM and dspy.Adapter will be key for DSPy 2.5 onwards. Old clients are unaffected for now.
This is a pretty significant release. New dspy.LM and dspy.Adapter will be key for DSPy 2.5 onwards. Old clients are unaffected for now.
Full Changelog: https://github.com/stanfordnlp/dspy/compare/2.4.15...2.4.16
Assertions patches by @arnavsinghvi11 in https://github.com/stanfordnlp/dspy/pull/1471
Full Changelog: https://github.com/stanfordnlp/dspy/compare/2.4.14...2.4.15
Nothing published for this version
supports configurable backoff time for LM/RM providers dspy.settings.configure(backoff_time = ...)
dspy.settings.configure(backoff_time = ...)Full Changelog: https://github.com/stanfordnlp/dspy/compare/v2.4.12...2.4.13
Supports treating dspy.Predict and dspy.ChainOfThought directly as Modules, e.g. can compile them without creating a wrapping dspy.Module object.
Supports treating dspy.Predict and dspy.ChainOfThought directly as Modules, e.g. can compile them without creating a wrapping dspy.Module object.
Improves the experimental=True Chat LM support from DSPy v2.4.11, which is quoted below for reference:
Initial support for new adapters, with far more smooth support for Chat LMs. The default behavior is unchanged from before. However, dspy.configure(experimental=True) is likely to lead to much more predictable and accurate zero-shot generations with Chat LMs. Tested against several variants of GPT-3.5, GPT-4, Llama3, Mixtral, and DBRX.
All sorts of PRs over the past two weeks.
All sorts of PRs over the past two weeks.
Most importantly, initial support for new adapters, with far more smooth support for Chat LMs.
The default behavior is unchanged from before. However, dspy.configure(experimental=True) is likely to lead to much more predictable and accurate zero-shot generations with Chat LMs. Tested against several variants of GPT-3.5, GPT-4, Llama3, Mixtral, and DBRX.
MIPROv2 and everything else from May 2024 and up to June 20th.
MIPROv2 and everything else from May 2024 and up to June 20th.
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →