diff --git a/src/precompile.jl b/src/precompile.jl
index 88fd48e..c7e071a 100644
--- a/src/precompile.jl
+++ b/src/precompile.jl
@@ -5,27 +5,16 @@ if Base.VERSION >= v"1.9"
         rng = MersenneTwister(7)
         A = randn(rng, 4, 3, 2)
         B = abs.(A)
-        A_med = randn(rng, 6, 5, 4)
 
         model = JoinModel(A, 2; geometry = :canonical)
         p0 = _solver_point(manifold(model), initial_point(model, :random; verbose = false))
-        model_med = JoinModel(A_med, 2; geometry = :canonical)
-        p0_med = _solver_point(
-            manifold(model_med),
-            initial_point(model_med, :random; verbose = false),
-        )
 
-        nn_model = JoinModel(B, 2; nonnegative = true, geometry = :softplus_metric)
-        p0_nn = _solver_point(
-            manifold(nn_model),
-            initial_point(nn_model, :random; verbose = false),
-        )
-
-        # The first LM call is dominated by JIT compilation in the operator-action
-        # path. Precompiling a minimal canonical and softplus NNCP workload moves
-        # that cost to package precompile time instead of first user execution.
+        # Keep this workload intentionally small: we want to cover the operator LM
+        # path and the public CP/NNCP frontends without forcing every user to pay
+        # a very large package precompile cost up front.
         @compile_workload begin
-            # Direct LM solve path on an already-prepared canonical CP model.
+            # Direct LM solve on a canonical CP model covers the core operator
+            # least-squares path without routing through the public frontend.
             solve(
                 LMSolver(),
                 model;
@@ -36,38 +25,8 @@ if Base.VERSION >= v"1.9"
                 verbose = false,
                 return_stats = true,
             )
-            solve(
-                LMSolver(),
-                model_med;
-                init = :random,
-                p0 = p0_med,
-                maxiter = 1,
-                tol = 1e-6,
-                verbose = false,
-                return_stats = true,
-            )
-            cpd(
-                A,
-                2;
-                solver = :lm,
-                init = :random,
-                p0 = p0,
-                maxiter = 1,
-                tol = 1e-6,
-                verbose = false,
-            )
-            cpd(
-                A_med,
-                2;
-                solver = :lm,
-                init = :random,
-                p0 = p0_med,
-                maxiter = 1,
-                tol = 1e-6,
-                verbose = false,
-            )
-            # Frontend LM route with ALS warm start, which is the path users hit
-            # most often when they request cpd(...; solver=:lm).
+
+            # Warm-started CPD is the most common public LM entry point.
             cpd(
                 A,
                 2;
@@ -78,26 +37,8 @@ if Base.VERSION >= v"1.9"
                 tol = 1e-6,
                 verbose = false,
             )
-            cpd(
-                A_med,
-                2;
-                solver = :lm,
-                init = :alswarm,
-                warm_steps = 2,
-                maxiter = 1,
-                tol = 1e-6,
-                verbose = false,
-            )
-            nncpd(
-                B,
-                2;
-                solver = :lm,
-                init = :random,
-                p0 = p0_nn,
-                maxiter = 1,
-                tol = 1e-6,
-                verbose = false,
-            )
+
+            # Nonnegative LM still goes through a distinct parameterization path.
             nncpd(
                 B,
                 2;