diff --git a/R/nearest_neighbor_data.R b/R/nearest_neighbor_data.R
index 7eaf36b40..a24051f61 100644
--- a/R/nearest_neighbor_data.R
+++ b/R/nearest_neighbor_data.R
@@ -93,9 +93,9 @@ set_pred(
     # model mode
     pre = function(x, object) {
       if (object$fit$response != "continuous") {
-        rlang::abort(
-          glue::glue("`kknn` model does not appear to use numeric predictions.",
-                     " Was the model fit with a continuous response variable?")
+        cli::cli_abort(
+          c("`kknn` model does not appear to use numeric predictions.",
+            "i" = "Was the model fit with a continuous response variable?")
         )
       }
       x
@@ -136,9 +136,9 @@ set_pred(
   value = list(
     pre = function(x, object) {
       if (!(object$fit$response %in% c("ordinal", "nominal"))) {
-        rlang::abort(
-          glue::glue("`kknn` model does not appear to use class predictions.",
-                     " Was the model fit with a factor response variable?")
+        cli::cli_abort(
+          c("`kknn` model does not appear to use class predictions.",
+            "i" = "Was the model fit with a factor response variable?")
         )
       }
       x
@@ -162,9 +162,9 @@ set_pred(
   value = list(
     pre = function(x, object) {
       if (!(object$fit$response %in% c("ordinal", "nominal"))) {
-        rlang::abort(
-          glue::glue("`kknn` model does not appear to use class predictions.",
-                     " Was the model fit with a factor response variable?")
+        cli::cli_abort(
+          c("`kknn` model does not appear to use class predictions.",
+            "i" = "Was the model fit with a factor response variable?")
         )
       }
       x
diff --git a/R/rand_forest.R b/R/rand_forest.R
index 2a6aed042..f15b5e906 100644
--- a/R/rand_forest.R
+++ b/R/rand_forest.R
@@ -99,9 +99,9 @@ translate.rand_forest <- function(x, engine = x$engine, ...) {
 
   if (x$engine == "spark") {
     if (x$mode == "unknown") {
-      rlang::abort(
-        glue::glue("For spark random forests models, the mode cannot ",
-                   "be 'unknown' if the specification is to be translated.")
+      cli::cli_abort(
+        "For spark random forest models, the mode cannot
+         be {.val unknown} if the specification is to be translated."
       )
     } else {
       arg_vals$type <- x$mode
@@ -121,7 +121,12 @@ translate.rand_forest <- function(x, engine = x$engine, ...) {
 
     if (any(names(arg_vals) == "importance")) {
       if (isTRUE(is.logical(quo_get_expr(arg_vals$importance)))) {
-        rlang::abort("`importance` should be a character value. See ?ranger::ranger.")
+        cli::cli_abort(
+          c(
+            "{.arg importance} should be a character value.",
+            "i" = "See ?ranger::ranger."
+          )
+        )
       }
     }
     # unless otherwise specified, classification models are probability forests
diff --git a/R/rand_forest_data.R b/R/rand_forest_data.R
index 5ea5e6f1d..e73406e78 100644
--- a/R/rand_forest_data.R
+++ b/R/rand_forest_data.R
@@ -64,11 +64,9 @@ ranger_confint <- function(object, new_data, ...) {
     if (object$fit$forest$treetype == "Probability estimation") {
       res <- ranger_class_confint(object, new_data, ...)
     } else {
-      rlang::abort(
-        glue::glue(
-          "Cannot compute confidence intervals for a ranger forest ",
-          "of type {object$fit$forest$treetype}."
-        )
+      cli::cli_abort(
+        "Cannot compute confidence intervals for a ranger forest
+         of type {.val {object$fit$forest$treetype}}."
       )
     }
   }
@@ -204,10 +202,10 @@ set_pred(
   value = list(
     pre = function(x, object) {
       if (object$fit$forest$treetype != "Probability estimation")
-        rlang::abort(
-          glue::glue(
-            "`ranger` model does not appear to use class probabilities. Was ",
-            "the model fit with `probability = TRUE`?"
+        cli::cli_abort(
+          c(
+            "`ranger` model does not appear to use class probabilities.",
+            "i" = "Was the model fit with `probability = TRUE`?"
           )
         )
       x
diff --git a/R/svm_linear.R b/R/svm_linear.R
index 2548cd499..29bac41a6 100644
--- a/R/svm_linear.R
+++ b/R/svm_linear.R
@@ -105,10 +105,9 @@ translate.svm_linear <- function(x, engine = x$engine, ...) {
         arg_vals$svr_eps <- 0.1
       if (!is_null(liblinear_type))
         if(!liblinear_type %in% 11:13)
-          rlang::abort(
-            paste0("The LiblineaR engine argument of `type` = ",
-                   liblinear_type,
-                   " does not correspond to an SVM regression model.")
+          cli::cli_abort(
+            "The LiblineaR engine argument {.code type = {liblinear_type}}
+             does not correspond to an SVM regression model."
           )
     } else if (x$mode == "classification") {
       if (!is_null(liblinear_type))
diff --git a/R/svm_rbf_data.R b/R/svm_rbf_data.R
index 84fad0cb4..5106ff4f1 100644
--- a/R/svm_rbf_data.R
+++ b/R/svm_rbf_data.R
@@ -296,9 +296,11 @@ set_pred(
   value = list(
     pre = function(x, object) {
       if (!object$fit$predict.prob) {
-        rlang::abort(
-          paste0("`svm` model does not appear to use class probabilities. Was ",
-                 "the model fit with `predict.prob = TRUE`?")
+        cli::cli_abort(
+          c(
+            "The model does not appear to use class probabilities.",
+            "i" = "Was the model fit with {.code predict.prob = TRUE}?"
+          )
         )
       }
       x