diff --git a/soroban-sdk/src/_migrating.rs b/soroban-sdk/src/_migrating.rs
index a298de57e..74f6c5b1b 100644
--- a/soroban-sdk/src/_migrating.rs
+++ b/soroban-sdk/src/_migrating.rs
@@ -8,7 +8,9 @@
 //! 2. Add support for [CAP-82: Checked 256-bit integer arithmetic host functions](https://github.com/stellar/stellar-protocol/blob/master/core/cap-0082.md).
 //!    New `checked_*` methods on [`U256`] and [`I256`] (`checked_add`, `checked_sub`,
 //!    `checked_mul`, `checked_div`, `checked_pow`, `checked_rem_euclid`, `checked_shl`,
-//!    `checked_shr`) return `Option` instead of panicking on overflow.
+//!    `checked_shr`) return `Option` instead of panicking on overflow. Also adds
+//!    `min_value` and `max_value` methods on [`U256`] and [`I256`] to fetch the
+//!    value bounds of each type.
 //!
 //! 3. Add support for [CAP-80: Host functions for efficient ZK BN254 use cases](https://github.com/stellar/stellar-protocol/blob/master/core/cap-0080.md).
 //!    [`BN254`] gains scalar field arithmetic (`Fr` `Add`/`Sub`/`Mul` traits, `pow`, `inv`),
diff --git a/soroban-sdk/src/num.rs b/soroban-sdk/src/num.rs
index fcdb39c00..e96b729af 100644
--- a/soroban-sdk/src/num.rs
+++ b/soroban-sdk/src/num.rs
@@ -205,6 +205,16 @@ impl_num_wrapping_val_type!(U256, U256Val, U256Small);
 impl U256 {
     pub const BITS: u32 = 256;
 
+    /// Returns the smallest value that can be represented by this type (0).
+    pub fn min_value(env: &Env) -> Self {
+        Self::from_u32(env, 0)
+    }
+
+    /// Returns the largest value that can be represented by this type (2^256 - 1).
+    pub fn max_value(env: &Env) -> Self {
+        Self::from_parts(env, u64::MAX, u64::MAX, u64::MAX, u64::MAX)
+    }
+
     fn is_zero(&self) -> bool {
         *self == U256::from_u32(&self.env, 0)
     }
@@ -438,6 +448,16 @@ impl_num_wrapping_val_type!(I256, I256Val, I256Small);
 impl I256 {
     pub const BITS: u32 = 256;
 
+    /// Returns the smallest value that can be represented by this type (-2^255).
+    pub fn min_value(env: &Env) -> Self {
+        Self::from_parts(env, i64::MIN, 0, 0, 0)
+    }
+
+    /// Returns the largest value that can be represented by this type (2^255 - 1).
+    pub fn max_value(env: &Env) -> Self {
+        Self::from_parts(env, i64::MAX, u64::MAX, u64::MAX, u64::MAX)
+    }
+
     fn is_zero(&self) -> bool {
         *self == I256::from_i32(&self.env, 0)
     }
@@ -446,11 +466,6 @@ impl I256 {
         *self == I256::from_i32(&self.env, -1)
     }
 
-    /// Returns the minimum value of I256 (-2^255).
-    pub fn min_value(env: &Env) -> Self {
-        I256::from_parts(env, i64::MIN, 0, 0, 0)
-    }
-
     pub fn from_i32(env: &Env, i: i32) -> Self {
         I256 {
             env: env.clone(),
@@ -847,6 +862,36 @@ mod test {
         assert_eq!(u3.rem_euclid(&u4), U256::from_u32(&env, 3));
     }
 
+    #[test]
+    fn test_u256_min() {
+        let env = Env::default();
+
+        let min = U256::min_value(&env);
+        assert_eq!(min, U256::from_u32(&env, 0));
+
+        let one = U256::from_u32(&env, 1);
+        assert_eq!(min.checked_sub(&one), None);
+        assert!(min.checked_add(&one).is_some());
+    }
+
+    #[test]
+    fn test_u256_max() {
+        let env = Env::default();
+
+        let max = U256::max_value(&env);
+        assert_eq!(
+            max,
+            U256::from_parts(&env, u64::MAX, u64::MAX, u64::MAX, u64::MAX)
+        );
+
+        let u128_max = U256::from_u128(&env, u128::MAX);
+        assert!(max > u128_max);
+
+        let one = U256::from_u32(&env, 1);
+        assert_eq!(max.checked_add(&one), None);
+        assert!(max.checked_sub(&one).is_some());
+    }
+
     #[test]
     fn test_u256_checked_arith() {
         let env = Env::default();
@@ -873,7 +918,7 @@ mod test {
         let zero = U256::from_u32(&env, 0);
         let one = U256::from_u32(&env, 1);
         let two = U256::from_u32(&env, 2);
-        let max = U256::from_parts(&env, u64::MAX, u64::MAX, u64::MAX, u64::MAX);
+        let max = U256::max_value(&env);
         assert_eq!(max.checked_add(&one), None);
         assert_eq!(zero.checked_sub(&one), None);
         assert_eq!(max.checked_mul(&two), None);
@@ -921,6 +966,39 @@ mod test {
         assert_eq!(u3.rem_euclid(&u4), I256::from_i32(&env, 1));
     }
 
+    #[test]
+    fn test_i256_min() {
+        let env = Env::default();
+
+        let min = I256::min_value(&env);
+        assert_eq!(min, I256::from_parts(&env, i64::MIN, 0, 0, 0));
+
+        let i128_min = I256::from_i128(&env, i128::MIN);
+        assert!(min < i128_min);
+
+        let one = I256::from_i32(&env, 1);
+        assert_eq!(min.checked_sub(&one), None);
+        assert!(min.checked_add(&one).is_some());
+    }
+
+    #[test]
+    fn test_i256_max() {
+        let env = Env::default();
+
+        let max = I256::max_value(&env);
+        assert_eq!(
+            max,
+            I256::from_parts(&env, i64::MAX, u64::MAX, u64::MAX, u64::MAX)
+        );
+
+        let i128_max = I256::from_i128(&env, i128::MAX);
+        assert!(max > i128_max);
+
+        let one = I256::from_i32(&env, 1);
+        assert_eq!(max.checked_add(&one), None);
+        assert!(max.checked_sub(&one).is_some());
+    }
+
     #[test]
     fn test_i256_checked_arith() {
         let env = Env::default();
@@ -948,8 +1026,8 @@ mod test {
         let one = I256::from_i32(&env, 1);
         let negative_one = I256::from_i32(&env, -1);
         let two = I256::from_i32(&env, 2);
-        let max = I256::from_parts(&env, i64::MAX, u64::MAX, u64::MAX, u64::MAX);
-        let min = I256::from_parts(&env, i64::MIN, 0, 0, 0);
+        let max = I256::max_value(&env);
+        let min = I256::min_value(&env);
         assert_eq!(max.checked_add(&one), None);
         assert_eq!(min.checked_sub(&one), None);
         assert_eq!(max.checked_mul(&two), None);