diff --git a/Cargo.lock b/Cargo.lock
index 7cfcd3a664..b205cea5d1 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -7,6 +7,7 @@ dependencies = [
  "criterion",
  "gc",
  "jemallocator",
+ "num-traits",
  "rand",
  "regex",
  "rustc-hash",
diff --git a/boa/Cargo.toml b/boa/Cargo.toml
index e8428876ed..dfd749beb7 100644
--- a/boa/Cargo.toml
+++ b/boa/Cargo.toml
@@ -14,6 +14,7 @@ edition = "2018"
 gc = { version = "0.3.4", features = ["derive"] }
 serde_json = "1.0.52"
 rand = "0.7.3"
+num-traits = "0.2.11"
 regex = "1.3.7"
 rustc-hash = "1.1.0"
 
diff --git a/boa/src/builtins/number/mod.rs b/boa/src/builtins/number/mod.rs
index dac495980e..d9e864ece5 100644
--- a/boa/src/builtins/number/mod.rs
+++ b/boa/src/builtins/number/mod.rs
@@ -23,6 +23,7 @@ use crate::{
     },
     exec::Interpreter,
 };
+use num_traits::float::FloatCore;
 use std::{borrow::Borrow, f64, ops::Deref};
 
 /// Helper function that converts a Value to a Number.
@@ -159,6 +160,129 @@ pub fn to_precision(this: &mut Value, args: &[Value], _ctx: &mut Interpreter) ->
     unimplemented!("TODO: Implement toPrecision");
 }
 
+const BUF_SIZE: usize = 2200;
+
+// https://golang.org/src/math/nextafter.go
+#[inline]
+fn next_after(x: f64, y: f64) -> f64 {
+    if x.is_nan() || y.is_nan() {
+        f64::NAN
+    } else if (x - y) == 0. {
+        x
+    } else if x == 0.0 {
+        f64::from_bits(1).copysign(y)
+    } else if y > x || x > 0.0 {
+        f64::from_bits(x.to_bits() + 1)
+    } else {
+        f64::from_bits(x.to_bits() - 1)
+    }
+}
+
+// https://chromium.googlesource.com/v8/v8/+/refs/heads/master/src/numbers/conversions.cc#1230
+pub fn num_to_string(mut value: f64, radix: u8) -> String {
+    assert!(radix >= 2);
+    assert!(radix <= 36);
+    assert!(value.is_finite());
+    // assert_ne!(0.0, value);
+
+    // Character array used for conversion.
+    // Temporary buffer for the result. We start with the decimal point in the
+    // middle and write to the left for the integer part and to the right for the
+    // fractional part. 1024 characters for the exponent and 52 for the mantissa
+    // either way, with additional space for sign, decimal point and string
+    // termination should be sufficient.
+    let mut buffer: [u8; BUF_SIZE] = [0; BUF_SIZE];
+    let (int_buf, frac_buf) = buffer.split_at_mut(BUF_SIZE / 2);
+    let mut fraction_cursor = 0;
+    let negative = value.is_sign_negative();
+    if negative {
+        value = -value
+    }
+    // Split the value into an integer part and a fractional part.
+    // let mut integer = value.trunc();
+    // let mut fraction = value.fract();
+    let mut integer = value.floor();
+    let mut fraction = value - integer;
+
+    // We only compute fractional digits up to the input double's precision.
+    let mut delta = 0.5 * (next_after(value, f64::MAX) - value);
+    delta = next_after(0.0, f64::MAX).max(delta);
+    assert!(delta > 0.0);
+    if fraction >= delta {
+        // Insert decimal point.
+        frac_buf[fraction_cursor] = b'.';
+        fraction_cursor += 1;
+        loop {
+            // Shift up by one digit.
+            fraction *= radix as f64;
+            delta *= radix as f64;
+            // Write digit.
+            let digit = fraction as u32;
+            frac_buf[fraction_cursor] = std::char::from_digit(digit, radix as u32).unwrap() as u8;
+            fraction_cursor += 1;
+            // Calculate remainder.
+            fraction -= digit as f64;
+            // Round to even.
+            if fraction + delta > 1.0
+                && (fraction > 0.5 || (fraction - 0.5) < f64::EPSILON && digit & 1 != 0)
+            {
+                loop {
+                    // We need to back trace already written digits in case of carry-over.
+                    fraction_cursor -= 1;
+                    if fraction_cursor == 0 {
+                        //              CHECK_EQ('.', buffer[fraction_cursor]);
+                        // Carry over to the integer part.
+                        integer += 1.;
+                        break;
+                    } else {
+                        let c: u8 = frac_buf[fraction_cursor];
+                        // Reconstruct digit.
+                        let digit_0 = (c as char).to_digit(10).unwrap();
+                        if digit_0 + 1 >= radix as u32 {
+                            continue;
+                        }
+                        frac_buf[fraction_cursor] =
+                            std::char::from_digit(digit_0 + 1, radix as u32).unwrap() as u8;
+                        fraction_cursor += 1;
+                        break;
+                    }
+                }
+                break;
+            }
+            if fraction < delta {
+                break;
+            }
+        }
+    }
+
+    // Compute integer digits. Fill unrepresented digits with zero.
+    let mut int_iter = int_buf.iter_mut().enumerate().rev(); //.rev();
+    while FloatCore::integer_decode(integer / f64::from(radix)).1 > 0 {
+        integer /= radix as f64;
+        *int_iter.next().unwrap().1 = b'0';
+    }
+
+    loop {
+        let remainder = integer % (radix as f64);
+        *int_iter.next().unwrap().1 =
+            std::char::from_digit(remainder as u32, radix as u32).unwrap() as u8;
+        integer = (integer - remainder) / radix as f64;
+        if integer <= 0f64 {
+            break;
+        }
+    }
+    // Add sign and terminate string.
+    if negative {
+        *int_iter.next().unwrap().1 = b'-';
+    }
+    assert!(fraction_cursor < BUF_SIZE);
+
+    let integer_cursor = int_iter.next().unwrap().0 + 1;
+    let fraction_cursor = fraction_cursor + BUF_SIZE / 2;
+    // dbg!("Number: {}, Radix: {}, Cursors: {}, {}", value, radix, integer_cursor, fraction_cursor);
+    String::from_utf8_lossy(&buffer[integer_cursor..fraction_cursor]).into()
+}
+
 /// `Number.prototype.toString( [radix] )`
 ///
 /// The `toString()` method returns a string representing the specified Number object.
@@ -169,8 +293,45 @@ pub fn to_precision(this: &mut Value, args: &[Value], _ctx: &mut Interpreter) ->
 ///
 /// [spec]: https://tc39.es/ecma262/#sec-number.prototype.tostring
 /// [mdn]: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Number/toString
-pub fn to_string(this: &mut Value, _args: &[Value], _ctx: &mut Interpreter) -> ResultValue {
-    Ok(Value::from(format!("{}", to_number(this).to_number())))
+pub fn to_string(this: &mut Value, args: &[Value], _ctx: &mut Interpreter) -> ResultValue {
+    // 1. Let x be ? thisNumberValue(this value).
+    let x = to_number(this).to_number();
+    // 2. If radix is undefined, let radixNumber be 10.
+    // 3. Else, let radixNumber be ? ToInteger(radix).
+    let radix_number = args.get(0).map_or(10, |arg| arg.to_integer()) as u8;
+
+    if x == -0. {
+        return Ok(Value::from("0"));
+    } else if x.is_nan() {
+        return Ok(Value::from("NaN"));
+    } else if x.is_infinite() && x.is_sign_positive() {
+        return Ok(Value::from("Infinity"));
+    } else if x.is_infinite() && x.is_sign_negative() {
+        return Ok(Value::from("-Infinity"));
+    }
+
+    // 4. If radixNumber < 2 or radixNumber > 36, throw a RangeError exception.
+    if radix_number < 2 || radix_number > 36 {
+        panic!("Radix must be between 2 and 36");
+    }
+
+    // 5. If radixNumber = 10, return ! ToString(x).
+    // This part should use exponential notations for long integer numbers commented tests
+    if radix_number == 10 {
+        // return Ok(to_value(format!("{}", to_number(this).to_num())));
+        return Ok(Value::from(format!("{}", x)));
+    }
+
+    // This is a Optimization from the v8 source code to print values that can fit in a single character
+    // Since the actual num_to_string allocates a 2200 bytes buffer for actual conversion
+    // I am not sure if this part is effective as the v8 equivalent https://chromium.googlesource.com/v8/v8/+/refs/heads/master/src/builtins/number.tq#53
+    // // Fast case where the result is a one character string.
+    // if x.is_sign_positive() && x.fract() == 0.0 && x < radix_number as f64 {
+    //     return Ok(to_value(format!("{}", std::char::from_digit(x as u32, radix_number as u32).unwrap())))
+    // }
+
+    // 6. Return the String representation of this Number value using the radix specified by radixNumber.
+    Ok(Value::from(num_to_string(x, radix_number)))
 }
 
 /// `Number.prototype.toString()`
diff --git a/boa/src/builtins/number/tests.rs b/boa/src/builtins/number/tests.rs
index fc9aea1cbc..e4cb7a4d2a 100644
--- a/boa/src/builtins/number/tests.rs
+++ b/boa/src/builtins/number/tests.rs
@@ -162,26 +162,210 @@ fn to_precision() {
 fn to_string() {
     let realm = Realm::create();
     let mut engine = Executor::new(realm);
-    let init = r#"
-        var default_string = Number().toString();
-        var int_string = Number(123).toString();
-        var float_string = Number(1.234).toString();
-        var exp_string = Number("1.2e+4").toString();
-        var neg_string = Number(-1.2).toString();
-        "#;
 
-    eprintln!("{}", forward(&mut engine, init));
-    let default_string = forward(&mut engine, "default_string");
-    let int_string = forward(&mut engine, "int_string");
-    let float_string = forward(&mut engine, "float_string");
-    let exp_string = forward(&mut engine, "exp_string");
-    let neg_string = forward(&mut engine, "neg_string");
-
-    assert_eq!(default_string, String::from("0"));
-    assert_eq!(int_string, String::from("123"));
-    assert_eq!(float_string, String::from("1.234"));
-    assert_eq!(exp_string, String::from("12000"));
-    assert_eq!(neg_string, String::from("-1.2"));
+    assert_eq!("NaN", &forward(&mut engine, "Number(NaN).toString()"));
+    assert_eq!("Infinity", &forward(&mut engine, "Number(1/0).toString()"));
+    assert_eq!(
+        "-Infinity",
+        &forward(&mut engine, "Number(-1/0).toString()")
+    );
+    assert_eq!("0", &forward(&mut engine, "Number(0).toString()"));
+    assert_eq!("9", &forward(&mut engine, "Number(9).toString()"));
+    assert_eq!("90", &forward(&mut engine, "Number(90).toString()"));
+    assert_eq!("90.12", &forward(&mut engine, "Number(90.12).toString()"));
+    assert_eq!("0.1", &forward(&mut engine, "Number(0.1).toString()"));
+    assert_eq!("0.01", &forward(&mut engine, "Number(0.01).toString()"));
+    assert_eq!("0.0123", &forward(&mut engine, "Number(0.0123).toString()"));
+    assert_eq!(
+        "0.00001",
+        &forward(&mut engine, "Number(0.00001).toString()")
+    );
+    assert_eq!(
+        "0.000001",
+        &forward(&mut engine, "Number(0.000001).toString()")
+    );
+    assert_eq!("NaN", &forward(&mut engine, "Number(NaN).toString(16)"));
+    assert_eq!(
+        "Infinity",
+        &forward(&mut engine, "Number(1/0).toString(16)")
+    );
+    assert_eq!(
+        "-Infinity",
+        &forward(&mut engine, "Number(-1/0).toString(16)")
+    );
+    assert_eq!("0", &forward(&mut engine, "Number(0).toString(16)"));
+    assert_eq!("9", &forward(&mut engine, "Number(9).toString(16)"));
+    assert_eq!("5a", &forward(&mut engine, "Number(90).toString(16)"));
+    assert_eq!(
+        "5a.1eb851eb852",
+        &forward(&mut engine, "Number(90.12).toString(16)")
+    );
+    assert_eq!(
+        "0.1999999999999a",
+        &forward(&mut engine, "Number(0.1).toString(16)")
+    );
+    assert_eq!(
+        "0.028f5c28f5c28f6",
+        &forward(&mut engine, "Number(0.01).toString(16)")
+    );
+    assert_eq!(
+        "0.032617c1bda511a",
+        &forward(&mut engine, "Number(0.0123).toString(16)")
+    );
+    assert_eq!(
+        "605f9f6dd18bc8000",
+        &forward(&mut engine, "Number(111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "3c3bc3a4a2f75c0000",
+        &forward(&mut engine, "Number(1111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "25a55a46e5da9a00000",
+        &forward(&mut engine, "Number(11111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "0.0000a7c5ac471b4788",
+        &forward(&mut engine, "Number(0.00001).toString(16)")
+    );
+    assert_eq!(
+        "0.000010c6f7a0b5ed8d",
+        &forward(&mut engine, "Number(0.000001).toString(16)")
+    );
+    assert_eq!(
+        "0.000001ad7f29abcaf48",
+        &forward(&mut engine, "Number(0.0000001).toString(16)")
+    );
+    assert_eq!(
+        "0.000002036565348d256",
+        &forward(&mut engine, "Number(0.00000012).toString(16)")
+    );
+    assert_eq!(
+        "0.0000021047ee22aa466",
+        &forward(&mut engine, "Number(0.000000123).toString(16)")
+    );
+    assert_eq!(
+        "0.0000002af31dc4611874",
+        &forward(&mut engine, "Number(0.00000001).toString(16)")
+    );
+    assert_eq!(
+        "0.000000338a23b87483be",
+        &forward(&mut engine, "Number(0.000000012).toString(16)")
+    );
+    assert_eq!(
+        "0.00000034d3fe36aaa0a2",
+        &forward(&mut engine, "Number(0.0000000123).toString(16)")
+    );
+
+    assert_eq!("0", &forward(&mut engine, "Number(-0).toString(16)"));
+    assert_eq!("-9", &forward(&mut engine, "Number(-9).toString(16)"));
+    assert_eq!("-5a", &forward(&mut engine, "Number(-90).toString(16)"));
+    assert_eq!(
+        "-5a.1eb851eb852",
+        &forward(&mut engine, "Number(-90.12).toString(16)")
+    );
+    assert_eq!(
+        "-0.1999999999999a",
+        &forward(&mut engine, "Number(-0.1).toString(16)")
+    );
+    assert_eq!(
+        "-0.028f5c28f5c28f6",
+        &forward(&mut engine, "Number(-0.01).toString(16)")
+    );
+    assert_eq!(
+        "-0.032617c1bda511a",
+        &forward(&mut engine, "Number(-0.0123).toString(16)")
+    );
+    assert_eq!(
+        "-605f9f6dd18bc8000",
+        &forward(&mut engine, "Number(-111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "-3c3bc3a4a2f75c0000",
+        &forward(&mut engine, "Number(-1111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "-25a55a46e5da9a00000",
+        &forward(&mut engine, "Number(-11111111111111111111111).toString(16)")
+    );
+    assert_eq!(
+        "-0.0000a7c5ac471b4788",
+        &forward(&mut engine, "Number(-0.00001).toString(16)")
+    );
+    assert_eq!(
+        "-0.000010c6f7a0b5ed8d",
+        &forward(&mut engine, "Number(-0.000001).toString(16)")
+    );
+    assert_eq!(
+        "-0.000001ad7f29abcaf48",
+        &forward(&mut engine, "Number(-0.0000001).toString(16)")
+    );
+    assert_eq!(
+        "-0.000002036565348d256",
+        &forward(&mut engine, "Number(-0.00000012).toString(16)")
+    );
+    assert_eq!(
+        "-0.0000021047ee22aa466",
+        &forward(&mut engine, "Number(-0.000000123).toString(16)")
+    );
+    assert_eq!(
+        "-0.0000002af31dc4611874",
+        &forward(&mut engine, "Number(-0.00000001).toString(16)")
+    );
+    assert_eq!(
+        "-0.000000338a23b87483be",
+        &forward(&mut engine, "Number(-0.000000012).toString(16)")
+    );
+    assert_eq!(
+        "-0.00000034d3fe36aaa0a2",
+        &forward(&mut engine, "Number(-0.0000000123).toString(16)")
+    );
+}
+
+#[test]
+#[ignore]
+// This tests fail for now since the Rust's default formatting for exponential format does not match the js spec.
+// https://github.com/jasonwilliams/boa/pull/381#discussion_r422458544
+fn num_to_string_exponential() {
+    let realm = Realm::create();
+    let mut engine = Executor::new(realm);
+
+    assert_eq!(
+        String::from("111111111111111110000"),
+        forward(&mut engine, "Number(111111111111111111111).toString()")
+    );
+    assert_eq!(
+        String::from("1.1111111111111111e+21"),
+        forward(&mut engine, "Number(1111111111111111111111).toString()")
+    );
+    assert_eq!(
+        String::from("1.1111111111111111e+22"),
+        forward(&mut engine, "Number(11111111111111111111111).toString()")
+    );
+    assert_eq!(
+        String::from("1e-7"),
+        forward(&mut engine, "Number(0.0000001).toString()")
+    );
+    assert_eq!(
+        String::from("1.2e-7"),
+        forward(&mut engine, "Number(0.00000012).toString()")
+    );
+    assert_eq!(
+        String::from("1.23e-7"),
+        forward(&mut engine, "Number(0.000000123).toString()")
+    );
+    assert_eq!(
+        String::from("1e-8"),
+        forward(&mut engine, "Number(0.00000001).toString()")
+    );
+    assert_eq!(
+        String::from("1.2e-8"),
+        forward(&mut engine, "Number(0.000000012).toString()")
+    );
+    assert_eq!(
+        String::from("1.23e-8"),
+        forward(&mut engine, "Number(0.0000000123).toString()")
+    );
 }
 
 #[test]