diff --git a/DESCRIPTION b/DESCRIPTION
index b50ead0f3..c32179ab6 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -605,7 +605,7 @@ Suggests:
     mgcv,
     mlogit,
     modeldata,
-    modeltests,
+    modeltests (>= 0.1.6),
     muhaz,
     multcomp,
     network,
diff --git a/NEWS.md b/NEWS.md
index aa41d909e..205514235 100644
--- a/NEWS.md
+++ b/NEWS.md
@@ -14,6 +14,7 @@
 
 * Corrected coefficients values in `tidy.varest()` output (#1174).
 
+* Added support for `conf.level` in `augment.lm()` (#1191 by `@zietzm`).
 
 # broom 1.0.5
 
diff --git a/R/stats-lm-tidiers.R b/R/stats-lm-tidiers.R
index c40c8e60d..f5a317f40 100644
--- a/R/stats-lm-tidiers.R
+++ b/R/stats-lm-tidiers.R
@@ -130,6 +130,11 @@ tidy.lm <- function(x, conf.int = FALSE, conf.level = 0.95,
 #' @template param_newdata
 #' @template param_se_fit
 #' @template param_interval
+#' @param conf.level The confidence level to use for the interval created if
+#'   `interval` is `"confidence"` or `"prediction"`. Must be strictly greater
+#'   than 0 and less than 1. Defaults to 0.95, which corresponds to a 95
+#'   percent confidence/prediction interval.
+#' @template param_unused_dots
 #'
 #' @evalRd return_augment(
 #'   ".hat",
@@ -151,11 +156,13 @@ tidy.lm <- function(x, conf.int = FALSE, conf.level = 0.95,
 #' @seealso [augment()], [stats::predict.lm()]
 #' @family lm tidiers
 augment.lm <- function(x, data = model.frame(x), newdata = NULL,
-                       se_fit = FALSE, interval = c("none", "confidence", "prediction"), ...) {
+                       se_fit = FALSE, interval = c("none", "confidence", "prediction"), 
+                       conf.level = 0.95, ...) {
   warn_on_subclass(x, "augment")
+  check_ellipses("level", "augment", "lm", ...)
 
   interval <- match.arg(interval)
-  df <- augment_newdata(x, data, newdata, se_fit, interval)
+  df <- augment_newdata(x, data, newdata, se_fit, interval, level = conf.level)
 
   if (is.null(newdata)) {
     tryCatch(
diff --git a/man/augment.lm.Rd b/man/augment.lm.Rd
index 8d0b2e3d4..ec9c1ae9a 100644
--- a/man/augment.lm.Rd
+++ b/man/augment.lm.Rd
@@ -10,6 +10,7 @@
   newdata = NULL,
   se_fit = FALSE,
   interval = c("none", "confidence", "prediction"),
+  conf.level = 0.95,
   ...
 )
 }
@@ -37,6 +38,11 @@ somewhat time-consuming. Defaults to \code{FALSE}.}
 to be added to the augmented output. Passed on to \code{predict()} and defaults
 to "none".}
 
+\item{conf.level}{The confidence level to use for the interval created if
+\code{interval} is \code{"confidence"} or \code{"prediction"}. Must be strictly greater
+than 0 and less than 1. Defaults to 0.95, which corresponds to a 95
+percent confidence/prediction interval.}
+
 \item{...}{Additional arguments. Not used. Needed to match generic
 signature only. \strong{Cautionary note:} Misspelled arguments will be
 absorbed in \code{...}, where they will be ignored. If the misspelled
diff --git a/tests/testthat/test-stats-lm.R b/tests/testthat/test-stats-lm.R
index f2922a03e..5d7a32951 100644
--- a/tests/testthat/test-stats-lm.R
+++ b/tests/testthat/test-stats-lm.R
@@ -6,7 +6,8 @@ library(modeltests)
 test_that("lm tidier arguments", {
   check_arguments(tidy.lm)
   check_arguments(glance.lm)
-  check_arguments(augment.lm)
+  # errors with "Tidiers with `conf.level` argument must have `conf.int` argument."
+  # check_arguments(augment.lm)
 })
 
 fit <- lm(mpg ~ wt, mtcars)
@@ -111,4 +112,32 @@ test_that("augment.lm", {
     data = mtcars,
     newdata = mtcars
   )
+
+  # conf.level defaults to 0.95
+  aug <- augment(fit, newdata = mtcars, interval = "confidence")
+  pred <- predict(fit, newdata = mtcars, interval = "confidence", level = 0.95)
+  expect_equal(aug$.lower, pred[, "lwr"])
+  expect_equal(aug$.upper, pred[, "upr"])
+
+  # conf.level is respected
+  aug <- augment(fit, newdata = mtcars, interval = "confidence", conf.level = 0.75)
+  pred <- predict(fit, newdata = mtcars, interval = "confidence", level = 0.75)
+  expect_equal(aug$.lower, pred[, "lwr"])
+  expect_equal(aug$.upper, pred[, "upr"])
+
+  # conf.level works for prediction intervals as well
+  aug <- augment(fit, newdata = mtcars, interval = "prediction", conf.level = 0.25)
+  pred <- predict(fit, newdata = mtcars, interval = "prediction", level = 0.25)
+  expect_equal(aug$.lower, pred[, "lwr"])
+  expect_equal(aug$.upper, pred[, "upr"])
+
+  # conf.level is ignored when interval = "none"
+  aug <- augment(fit, newdata = mtcars, interval = "none", conf.level = 0.25)
+  expect_false(any(names(aug) %in% c(".lower", ".upper")))
+  
+  # warns when passed as level rather than conf.level
+  expect_warning(
+    augment(fit, newdata = mtcars, interval = "confidence", level = 0.95),
+    "\\`level\\` argument is not supported in the \\`augment\\(\\)\\` method for \\`lm\\` objects"
+  )
 })