diff --git a/Project.toml b/Project.toml
index aeb9ce16a2a..89020d7b321 100644
--- a/Project.toml
+++ b/Project.toml
@@ -1,6 +1,6 @@
 name = "JuMP"
 uuid = "4076af6c-e467-56ae-b986-b466b2749572"
-version = "1.30.1"
+version = "1.31.0"
 repo = "https://github.com/jump-dev/JuMP.jl.git"
 
 [deps]
diff --git a/docs/Project.toml b/docs/Project.toml
index 43a00daffce..841d6cd530c 100644
--- a/docs/Project.toml
+++ b/docs/Project.toml
@@ -54,17 +54,17 @@ Clarabel = "=0.11.1"
 Clustering = "0.15.8"
 DSP = "0.8.5"
 DataFrames = "1.8.2"
-DifferentiationInterface = "0.7.18"
+DifferentiationInterface = "0.7.20"
 DimensionalData = "=0.30.1"
 Distributions = "0.25.129"
 Documenter = "=1.17.0"
 DocumenterCitations = "1.4.1"
 Dualization = "0.7.7"
-Enzyme = "=0.13.173"
+Enzyme = "=0.13.187"
 ForwardDiff = "1.4.1"
 Gurobi = "=1.9.2"
 HTTP = "1.11.0"
-HiGHS = "=1.24.0"
+HiGHS = "=1.24.1"
 Images = "0.26.2"
 Interpolations = "0.16.3"
 Ipopt = "=1.15.0"
@@ -74,13 +74,13 @@ LinearOperatorCollection = "2.4.2"
 Literate = "2.21.0"
 MarkdownAST = "0.1.3"
 MathOptChordalDecomposition = "=0.2.0"
-MathOptInterface = "=1.51.1"
-MultiObjectiveAlgorithms = "=1.11.0"
+MathOptInterface = "=1.51.2"
+MultiObjectiveAlgorithms = "=1.12.0"
 PATHSolver = "=1.7.9"
 ParametricOptInterface = "0.15.3"
 Plots = "1.41.6"
 RegularizedLeastSquares = "0.16.12"
-SCS = "=2.6.3"
+SCS = "=2.6.4"
 SQLite = "1.8.1"
 SpecialFunctions = "2.8.0"
 StatsPlots = "0.15.8"
diff --git a/docs/packages.toml b/docs/packages.toml
index 21c6a1099a6..6fe418ed857 100644
--- a/docs/packages.toml
+++ b/docs/packages.toml
@@ -63,14 +63,14 @@
 [Gurobi]
     rev = "v1.9.2"
 [HiGHS]
-    rev = "v1.24.0"
+    rev = "v1.24.1"
 [Hypatia]
     rev = "v0.10.0"
     has_html = true
 [Ipopt]
     rev = "v1.15.0"
 [KNITRO]
-    rev = "v1.2.1"
+    rev = "v1.3.0"
 [MathOptAnalyzer]
     rev = "v0.1.2"
 [MathOptIIS]
@@ -78,18 +78,18 @@
 [MathOptLazy]
     rev = "v0.1.0"
 [MiniZinc]
-    rev = "v0.3.14"
+    rev = "v0.3.15"
 [MosekTools]
     rev = "v0.15.10"
 [MultiObjectiveAlgorithms]
-    rev = "v1.11.0"
+    rev = "v1.12.0"
     has_html = true
 [NEOSServer]
     rev = "v1.2.0"
 [NLopt]
     rev = "v1.2.1"
 [Pajarito]
-    rev = "4c7efa915bdc900b51d1c8290dc771403b11230a"
+    rev = "v0.8.3"
 [ParametricOptInterface]
     rev = "v0.15.3"
     extension = true
@@ -105,7 +105,7 @@
     rev = "v0.8.1"
     extension = true
 [SCS]
-    rev = "v2.6.3"
+    rev = "v2.6.4"
 [SDPA]
     rev = "v0.6.1"
 [SDPLR]
@@ -120,7 +120,7 @@
     rev = "v0.8.0"
     extension = true
 [Xpress]
-    rev = "v0.18.0"
+    rev = "v0.18.1"
 
 # =========================== non-jump-dev packages ========================== #
 
@@ -155,11 +155,11 @@
     rev = "v1.1.33"
 [COSMO]
     user = "oxfordcontrol"
-    rev = "v0.8.10"
+    rev = "v0.8.11"
     has_html = true
 [CuPDLPx]
     user = "MIT-Lu-Lab"
-    rev = "v0.2.9"
+    rev = "v0.3.0"
 [DAQP]
     user = "darnstrom"
     rev = "v0.8.0"
@@ -198,7 +198,7 @@
     rev  = "v0.2.2"
 [MathOptAI]
     user = "lanl-ansi"
-    rev = "v0.2.5"
+    rev = "v0.2.6"
     extension = true
 [MathOptChordalDecomposition]
     user = "samuelsonric"
diff --git a/docs/src/changelog.md b/docs/src/changelog.md
index c2044441e35..eddd275897e 100644
--- a/docs/src/changelog.md
+++ b/docs/src/changelog.md
@@ -7,6 +7,59 @@ CurrentModule = JuMP
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## Version v1.31.0 (July 21, 2026)
+
+### Added
+
+ - As part of our long term goal to support vector-valued nonlinear expressions,
+   we added experimental support for `AbstractArray` in nonlinear expressions.
+   You can now build nonlinear expressions containing arrays, although there is
+   no corresponding support for this via solvers like Ipopt. We hope to change
+   this in future releases. (#3451)
+ - Common subexpressions are now cached when building
+   `MOI.ScalarNonlinearExpression`. This can result in a large speed-up for some
+   models, but we are also concerned about the potential for increased memory
+   usage in some models. If you notice changes (positive or negative), please
+   tell us via a GitHub issue. Related to this change, [`moi_function`](@ref)
+   now accepts a `model` as the first argument. Existing single-argument methods
+   will still work, but future callers should switch to the two-argument
+   version. (#4032), (#4189), (#4190), (#4200)
+ - We added various features to help AI agents understand JuMP (and also help
+   non-AI-assisted humans):
+    - We improved the `JuMP` module docstring so that `? JuMP` now describes
+      where to find the documentation and how to ask for help (#4174)
+    - We rewrote error messages to be clearer and include an actionable next
+      step. This may break tests if you relied on checking the exact string of
+      an error message (#4172), (#4176), (#4177), (#4178), (#4180), (#4182)
+    - We added learning intentions to all of the tutorials in the documentation
+      (#4170)
+    - We added an `AGENTS.md` (and `CLAUDE.md`), and added requests in multiple
+      parts of the documentation for users to report issues to the maintainers
+      (#4184), (#4185)
+
+### Fixed
+
+ - Fixed a bug with `BridgeableConstraint` and `model_convert` (#4183)
+ - Fixed a failing test on Julia nightly (#4173)
+ - Fixed a bug hashing DenseAxisArray (#4192)
+ - Fixed a bug so that we maintain the order of expression arguments when
+   calling [`flatten!`](@ref) (#4196)
+ - Fixed a bug where `0.0` is added to `GenericNonlinearExpr` in operators like
+   `sum` (#4197)
+ - Fixed the error messages when Hermitian variables are passed invalid data to
+   explain what data was incorrect (#4198)
+ - Fixed checking if problem is an LP in [`lp_sensitivity_report`](@ref) (#4199)
+ - Removed some unnecessary parentheses when printing nonlinear expressions (#4201)
+
+### Other
+
+ - Updated various parts of the documentation (#4158), (#4159), (#4164),
+   (#4165), (#4167), (#4179), (#4186), (#4187), (#4193), (#4195)
+ - Re-enabled `depwarn=error` in tests (#4161)
+ - Added a new [Sport scheduling](@ref) tutorial (#4160) (#4162)
+ - Updated various GitHub action versions (#4168), (#4181)
+ - Fixed formatting for new JuliaFormatter release (#4169)
+
 ## Version v1.30.1 (April 24, 2026)
 
 ### Fixed
diff --git a/docs/styles/config/vocabularies/JuMP/accept.txt b/docs/styles/config/vocabularies/JuMP/accept.txt
index 42173089018..417e57a053c 100644
--- a/docs/styles/config/vocabularies/JuMP/accept.txt
+++ b/docs/styles/config/vocabularies/JuMP/accept.txt
@@ -63,6 +63,7 @@ sparsifying
 subexpression(?s)
 subgraph(?s)
 sublicense
+subsolver
 subtyping
 SVM(?s)
 sysimage
@@ -141,6 +142,7 @@ Artelys
 (Cbc|CBC)
 (cdd|CDD)
 (cddlib|CDDLib)
+Chuffed
 Clarabel
 Clp
 CPLEXLink