Befunge-93, 47 bytes

86*:77*" ,4"5+:" ,wef a ssim ,5"3-" ,"77*>:#,_@

Try it online!

Explanation

Pushes the ASCII values of the string reversed (due to the nature of the stack).

Breakdown

86*                                             Push 48, the ASCII value of 0
   :                                            Duplicate
    77*                                         Push 49, the ASCII value of 1
       " ,4"5+                                  Push ASCII values of ` ,4", then add the character `4` by 5 to make it `9`
              :                                 Duplicate
               " ,wef a ssim ,5"3-              Push ASCII values of ` ,wef a ssim ,5`, then subtract character `5` by 3 to make it 2
                                  " ,"          Push ASCII values of ` ,`
                                      77*       Push 49, the ASCII value of 1
                                         >:#,_@ Print all the ASCII characters in the stack

Perl 5 + -M5.10.0, 31 bytes

Bit of a boring answer, calls say using (stringwise) bitwise negation operator (~) on the bitflipped string.

say~"........................."

Try it online!

Uiua 0.11.0, 29 bytes

-3"4/#5/#plvv#d#ihz/#<</#433"

Explanation:
  "4/#5/#plvv#d#ihz/#<</#433" # A string 
-3                            # Shift each character to the left by 3 (in Unicode)

See it in action

C#, 68 bytes

System.Console.Write($"{4-3}, {5-3}, miss a few, {33*3}, {64*4:X}");

YASEPL, 60 54 chars

=a)commaſspace=b$48+›#a+›#a#"miss a few"#a+7››#a-8›-››

if you're wondering, this is a pretty stupid esolang I made a while ago in Node.JS

GNU dc, 41 bytes

Kzn[, ]nzzn[, miss a few, ]n33z*dn[, ]n+n

Try it online!

Perl 5, 42 bytes

$,=", ";say!!3,5-3,"miss a few",3*33,4*5*5

Sets the record separator, and prints a few numbers and a string.

Try it online!

Perl 5, 43 bytes

say"C, D, miss a few, KK, CBB"=~y,A-L,/-:,r

Translates B-K to 0-9 (extended by one more ASCII code on each side of the range) and prints.

Try it online!

Perl 5, 44 bytes

$,=", ";say unpack"c3/acc","

miss a fewcd"

Sets the record separator, and unpacks values to be printed.

Try it online!

Knight, 50 bytes

O+++++++''-6 5', '-6 4', miss a few, 'A'c'', 'A'd'

Try It Online!

Probably not optimal but whatever.

><>, 44 bytes

68*:77*' ,>'5-:' ,wef a ssim ,':6+' ,':5+>o<

Try it online!

Java (JDK), 52 bytes

i->4-3+", "+(5-3)+", miss a few, "+(33*3+", "+5*5*4)

Try it online!

MathGolf, 23 22 bytes

)└"miss a few"♀┐\]û, u

Try it online.

Explanation:

)                      # +1 (since the stack is empty, it'll use an implicit 0)
 └                     # Push the top of the stack + 1 (without popping)
  "miss a few"         # Push string "miss a few"
              ♀        # Push 100
               ┐       # Push the top of the stack - 1 (without popping)
                \      # Swap the top two values on the stack
                 ]     # Wrap all values on the stack into a list
                  û,   # Push string ", "
                     u # Join the list with this ", " as delimiter
                       # (after which the entire stack is output implicitly as result)

Even if we were allowed to use 0129, the byte-count would remain the same (the )└ could simply be 12).

!@#$%^&*()_+, 34 bytes

^!# ,@@^# ,wef a ssim ,(@)c# ,@@d#

Try it online!

Explaination:

^!# ,@@^# ,wef a ssim ,(@)c# ,@@d# (implicit push 0)
^                                  Top of stack += 1
 !#                                Output Top of stack as number without popping
    ,@@                            Output ", "
       ^                           Top of stack += 1
        #                          Output Top of stack as number
          ,wef a ssim ,(@)         Output ", miss a few, "
                          c#       Push 99 onto the stack and output
                             ,@@   output ", "
                                d# Push 100 onto the stack and output

Perl 6, 45 bytes

say join ", ",5-4,5-3,"miss a few",3*33,4*5*5

Try it online!

Pip, 23 bytes

[o#k"miss a few"h+vh]Jk

Try it online!

having a builtin for , is sadly not enough

-3 bytes from DLosc.

CJam, 28 26 25 bytes

Two bytes off thanks to @Lyxal, and one off thanks to @Lynn!

XY"miss a fewc")i_)]", "*

Try it online!

Explanation

X              e# Push contents of variable X (predefined value): 1
Y              e# Push contents of variable Y (predefined value): 2
"miss a fewc"  e# Push this string
)              e# Detach the last character from that string
i              e# Convert to integer: gives 99
_)             e# Duplicate, increment: gives 100
]              e# Concatenate stack contents into an array
", "           e# Push this string 
*              e# Join the array entries with the specified string
               e# Implicit display 

Haskell, 54 bytes

p=pred
main=putStr$p.p.p<$>"4/#5/#plvv#d#ihz/#<</#433"

Try it online!

Japt, 22 bytes

Cì p`ÚÍ a few`#cL qSi,

Test it

Cì p`...`#cL qSi,
C                     :12
 ì                    :To digit array
   p                  :Push
    `...`             :  Compressed string "miss a few"
         #c           :  Charcode of "c" (99)
           L          :  100
             q        :Join with
              S       :  Space
               i,     :  Prepended with a comma

Python 3, 56 bytes

lambda:f"{3//3}, {8//4}, miss a few, {3*4*8+3}, {4*5*5}"

Try it online! Uses basic math and an f-string to get the job done.

AWK, 49 bytes

END{print++x", "5-3", miss a few, "3*33", "4*5*5}

Assuming empty input (otherwise change END to BEGIN for a 51-byte answer).

Try it online!

Deadfish~, 187 178 bytes

{iiiii}dcdddddc{d}ddc{ii}ddcddddddc{d}ddc{{i}dd}dddcddddc{i}cc{{d}ii}dddc{{i}dddd}iiiiic{{d}iiii}dddddc{{i}ddd}cdc{i}{i}ddc{{d}iii}dddddc{d}ddc{ii}iiiiicc{d}dddc{d}ddc{ii}dddcdcc

The previous version was bad.

Try it online!

Scratch, 110 bytes

Try it online!

Easy sauce. Alternatively, 12 blocks.

when gf clicked
say(join((4)-(3))(join[, ](join((5)-(3))(join[, miss a few, ](join((3)*(33))(join[, ]((66)+(34

PHP, 69 bytes

<?=55-54,', ',floor((5*4)/7),', ','miss a few ',64746/654,', ',5*5*4;

Barrel, 31 34 bytes

+n, +n, 'miss a few, '#44+#45+n, +n

This is my own esolang, follow the link for more info.

Increments the accumulator, prints it, implicitly prints some strings (, and miss a few, are implicitly printed), does a loop on accumulator incrementation, and finishes as it started.

Edit: I realized that "97" used a "9". smacks forehead

Arn, 22 bytes

˜▒çc>åžî¼#(¸“├)øW§Ò┴Ì"

Try it!

Explained

Just the string literal. Pretty boring, the next-shortest I could get without this was 29 bytes.

Vyxal, 21 20 19 bytes

3ɽ÷«eeȮǒḋp«₁‹₁W‛, j

Try it Online!

Explained

3ɽ÷      # stack = [1, 2]
«eeȮǒḋp« # stack = [1, 2, "miss a few"] 
₁‹₁      # stack = [1, 2, "miss a few", 99, 100]
W‛, j    # join stack on ", "

GolfScript, 34 bytes

!.)', ':c.@\'miss a few'3c\33*.c\)

Try it online!

!                                   Push 1 to the stack.
 .)                                 Copy the 1 and increment.
   ',':c                            Define c as ', ' to save characters when commas are needed.
        .@\                         Copy the ', ' and move stuff around into the right order.
           'miss a few'             Push the string to the stack.
                       3c\          Push 3 and ', ' then swap them.
                          33*       Multiply 3*33 to get 99.
                             .c\)   Copy the 99, push ', ' then swap them and increment the last 99.

APL (Dyalog Unicode), 45 bytes

(5↓7⌽∊'miss a few, ',⍨(', ',⍨⍕)¨5↑(3-5)⌽⍳77+!4)

 5↓ ⍝ Drop the first 5 elements
   7⌽ ⍝ Rotate left 7
     ∊'miss a few, ', ⍝ Concatenate and enlist
       ⍨ ⍝ Swap arguments
         (', ',⍨⍕)¨ ⍝ Encode as character and concatenate after swapping arguments
           5↑ ⍝ Take the first 5 elements
             (3-5)⌽ ⍝ Rotate right 2
               ⍳77+!4 ⍝ Create range from 0–100 inclusive

Try it online!

Javascript, 54 bytes

[`MSwgMi`,`OTksIDEwMA`].map(atob).join`, miss a few, `

*inspired by @arnauld 's answer

Pxem, Filename: 32 bytes + Content: 4 bytes = 36 bytes, depends on ASCII-compatible.

• Filename (escaped unprintables): \001.n.e\002.n.emiss a few.p.ec.n.ed.n
• Content: , .p

Try it online!

How it works

• Pxem is a stack-based procedure language.
• .n is printf("%d",pop) unless empty.
• .e calls subroutine (which is content).
• .p is putc pop until empty.
• Other non-command substrings are literal: pushed from backward.

CSASM v2.1.1, 145 bytes

func main:
push 4
push 3
sub
push ", "
add
push 5
push 3
sub
add
push ", miss a few, "
add
push 33
push 3
mul
add
push ", "
add
push 5
dup
mul
push 4
mul
add
print
ret
end

Commented:

func main:
    ; Push 4 - 3
    push 4
    push 3
    sub
    ; String concatenation:  1 + ", "
    push ", "
    add
    ; Push 5 - 3
    push 5
    push 3
    sub
    ; String concatenation:  "1, " + 2
    add
    push ", miss a few, "
    ; String concatenation:  "1, 2" + ", miss a few, "
    add
    ; Push 33 * 3
    push 33
    push 3
    mul
    ; String concatenation:  "1, 2, miss a few, " + 99
    add
    push ", "
    ; String concatenation:  "1, 2, miss a few, 99" + ", "
    add
    ; Push a 5 and multiply it by itself
    push 5
    dup
    mul
    ; Push 25 * 4
    push 4
    mul
    ; String concatenation:  "1, 2, miss a few, 99, " + 100
    add
    ; Print the value on the top of the stack:
    ;   Stack: [ "1, 2, miss a few, 99, 100" ]
    print
    ret
end

factor, 66 61 bytes

Shaved 5 bytes thanks to @Bubbler!

4 3 - 5 3 - 33 3 * CHAR: d "%d, %d, miss a few, %d, %d"printf

Try it online!

Original version:

4 3 - 5 3 - 33 3 * dup 3 - 4 + "%d, %d, miss a few, %d, %d" printf

My first factor line. Probably can be shorter…

Try it online!

x86 opcode, 42 41 bytes

BE 8F 81 04 08 29 C9 B1 | 1A AC 04 6E AA E2 FA C3
BE B2 C4 BE B2 FF FB 05 | 05 B2 F3 B2 F8 F7 09 BE
B2 CB CB BE B2 C3 C2 C2 | 92 

fun:	mov esi, src
	sub ecx, ecx
	mov cl, srcEnd-src
x:	lodsb
	add al, '1'-0xC3
	stosb
	loop x
src:	ret
	db ', 2, miss a few, 99, 100', 0
srcEnd:
repeat srcEnd-src-1
	load a byte from src+%
	store byte (a-'1'+0xC3) and 0xFF at src+%
end repeat
funEnd:

Try it online!

Function that output to [esi] assuming enough space, should be placed on right place such that src contain no invalid chars

AWK, 53 bytes

$a=++x", "++x", miss a few, "(z=(++y y)-x)z", "++z--y

Try it online!

This abuses a couple of AWK language features, including that all variables are initialized to blank/0/False (context sensitive), that it will happily convert strings to numeric values to make an arithmetic operator work, and it will also convert numeric values to strings in order to append characters to it.

First it gets 1, by incrementing variable x

++x", "

then 2, miss a few, just does the same thing again...

++x", miss a few, "

getting a 9 is a bit more work,

(++y y)-x

which works by getting a 1 with ++y then appended it to itself to get 11 then subtracting 2 with -x.

Using that to set the variable z means we can appended another 9 on the end with,

(z=...)z

That just leaves the last bit , 100 which we can get by incrementing the 9 and appending to that 1 decremented.

", "++z--y

I'm pretty sure there are other ways to do this in AWK that are shorter, but I like that this doesn't use any numbers.

One last thing I forgo to mention... The string is printed since,

$a=...

is interpreted as $0=... which evaluates to True, causing AWK print the value of $0, which was just set to the output we want.

BRASCA, 41 36 bytes

`<0x01>, <0x02>, miss a few, `I,[o]H{n` ,`ooHn

Replace <0x01> and <0x02> with ascii codepoints 1 and 2.

Try it!

Rust, 68 Bytes

fn main(){print!("{}, {}, miss a few, {}, {}",4-3,5-3,66+33,63+37)}

This is my first attempt at code golf Yahoo! Try it online

Rust, 56 bytes

Cutting out the main{} on tio saves 12 bytes

print!("{}, {}, miss a few, {}, {}",4-3,5-3,66+33,63+37)

Try it online!

PHP, 30 bytes

This contains unprintables so the link is to a Bash program that reconstructs the file and validates it.

<?=~.........................;

Try it online!

Explanation

prints the bitflipped (~) string 1, 2, miss a few, 99, 100.

Javascript (ES6), 70 bytes

Old 75:

a=8-7;b=8-6;c=8-8;d=4+5;_=>`${a}, ${b}, miss a few, ${d}${d}, ${a}${c}${c}`

New 50:

_=>[8-7,8-6,'miss a few',(4+5)*(5+5)+4+5,(8-7)*(55+55-5-5)].join(', ')

by far not the best js answer, ill widdle.

Templates Considered Harmful, 129 bytes

Cat<Cat<Cat<Cat<Cat<T,St<44,' '>>,Add<T,T>>,St<44,' ','m','i','s','s',' ','a',' ','f','e','w',44,' ',57,57,44,' '>>,T>,St<48,48>>

Try it online!

Tries to represent each character as its ascii value. If that contains a forbidden number, replace it with its char literal. This only leaves 1 and 2 to deal with.

1 is replaced with T, which evaluates to I<1> which is cast into string 1 when concatenated.
2 is replaced with Add<T,T>, which evaluates to I<2> which is cast into string 2 when concatenated.

Cat<
 ...
     T,                                                                    # "1"
     St<44,' '>>,                                                          # ", "
    Add<T,T>>,                                                             # "2"
   St<44,' ','m','i','s','s',' ','a',' ','f','e','w',44,' ',57,57,44,' '>>,# ", miss a few, 99, "
  T>,                                                                      # "1"
 St<48,48>>                                                                # "00"

C++17

#include <iostream>
main()
{
  std::cout << 4 - 3 << ", " << 5 - 3 << ", miss a few, " << 33 * 3 << ", " << 5 * 5 * 4 << "\n";
}

Elixir, 58 49 bytes

[?b-?a,?c-?a,"miss a few",?c,?d]|>Enum.join(", ")

Try it online!

x86_16 machine code, 146 bytes

Thank's to @mik for tips and trick :))

Replace : add al, "0" / add dl, "0" with add al, 2fh + inc al and add dl, 2fh + inc dl to avoid "0" in the source code.

B3 04          MOV BL, 4
B1 03          MOV CL, 3
2A D9          SUB BL, CL
B4 0E          MOV AH, 0EH
8A C3          MOV AL, BL
04 2F          ADD AL, 2FH
FE C0          INC AL
CD 10          INT 10H
B0 2C          MOV AL, ","
CD 10          INT 10H
80 C3 2F       ADD BL, 2FH
FE C3          INC BL
FE C3          INC BL
8A C3          MOV AL, BL
CD 10          INT 10H
B8 2C 0E       MOV AX, 0E2CH
CD 10          INT 10H
B4 09          MOV AH, 09h
BA 7A 01       MOV DX, OFFSET TXT
CD 21          INT 21H
B8 2C 0E       MOV AX, 0E2CH
CD 10          INT 10H
                                      
B8 20 0E       MOV AX, 0E20H
CD 10          INT 10H
               
B0 21          MOV AL, 33
B3 03          MOV BL, 3
F6 E3          MUL BL
B3 0A          MOV BL, 10
33 C9          XOR CX, CX
           LOOP1:
33 D2          XOR DX, DX
F7 F3          DIV BX
52             PUSH DX
41             INC CX
85 C0          TEST AX, AX
75 F6          JNZ LOOP1
    PRINT_DIGIT1:
5A             POP DX
80 C2 2F       ADD DL, 2FH
FE C2          INC DL
B4 02          MOV AH, 02H
CD 21          INT 21H
E2 F6          LOOP PRINT_DIGIT1
                                      
B8 2C 0E       MOV AX, 0E2CH
CD 10          INT 10H
B8 20 0E       MOV AX, 0E20H
CD 10          INT 10H
                                      
B0 04          MOV AL, 4
B3 05          MOV BL, 5
F6 E3          MUL BL
F6 E3          MUL BL
B3 0A          MOV BL, 10
33 C9          XOR CX, CX
           LOOP2:
33 D2          XOR DX, DX
F7 F3          DIV BX
52             PUSH DX
41             INC CX
85 C0          TEST AX, AX
75 F6          JNZ LOOP2
    PRINT_DIGIT2:
5A             POP DX
80 C2 2F       ADD DL, 2FH
FE C2          INC DL
B4 02          MOV AH, 02H
CD 21          INT 21H
E2 F6          LOOP PRINT_DIGIT2
            EXIT:
B8 00 4C       MOV AX, 4C00H
CD 21          INT 21H
                                                                              
20 6D 69 73 73 20 61 20 66 65 77 24   TXT DB " miss a few$"

Run is DOSBox :

C (gcc), 54 56 bytes

l(){printf("%c, %d, miss a few %d, %o",7*7,5-3,'c',64);}

Try it online!

Let's save a couple bytes using Octal for the 100. Thanks to Sheik Yerbouti for saving a byte and point out the no commas issue.

Yabasic, 46 bytes

?5-4,", ",5-3,", miss a few, ",33*3,", ",5*5*4

Try it online!

Replaced calculated numbers for 1 and 2 with \x hex codes for ASCII 49 and 50 and saved 3 bytes!

(Can't do the above because it results in \x31 and \x32 which of course contain 1 and 2...)

C++, 56 bytes

[]{for(int x:"4/#5/#plvv#d#ihz/#<</#433"s)putchar(x-3);}

Inspired by a comment.

Usage:

#include <string>
#include <stdio.h>

using namespace std;

int main() {
    auto k = [](){for(int x:"4/#5/#plvv#d#ihz/#<</#433"s)putchar(x-3);};
    k();
    return 0;
}

Try it online!

PHP, 58 54 bytes

<?=join(', ',[$i=3/3,++$i,'miss a few',$i=33*3,++$i]);

PHP doesnt really want to be typed short, but I liked coming up with this :)

PHP, 57 56 53 bytes

printf('%c, %d, miss a few, %d, %o',7*7,5-3,33*3,64);

Try it online!

A printf, some spices, stir

EDIT: -1 byte thanks to Digital Trauma

EDIT 2: -3 bytes thanks to Michael Dorgan for octal, thought about 153 with %x for hex but it has a 1..

Also, less fun but 2 bytes shorter by manatwork, improved by BadHorsie (it seems that <?= can also take multiple arguments with comas):

PHP, 51 49 bytes

<?=5-4,', ',5-3,', miss a few, ',33*3,', ',5*5*4;

Try it online!

Stax, 20 19 17 bytes

saved 2 bytes thanks to Razetime

ü←φr5╧wL'•↔♦$▒ò╠‼

Run and debug it

Original brain-dead solution for reference purposes only:

Stax, 22 bytes

ü2Φÿßbÿ»f╙m¡6¿₧|░╘;┬u»

Run and debug it

Charcoal, 21 bytes

”'3?Sxv7´Ｄ｜λ*§ξ¬Ｅ‹Ｄαg

Try it online! Compression again. Best I could do without compression is 23 bytes:

⪫⟦ＬψＬχmiss a few℅c℅d⟧, 

Try it online! Link is to verbose version of code.

Python 3, 54, 51, 47 42 bytes

Saved 2 bytes thanks to Digial Trauma's suggestion to replace 8*8+35 = 99 with 33*3 = 99.

Saved 1 byte by realizing that I could replace 8*8+36 = 100 with 4*5*5 = 100.

Saved 4 bytes thanks to dingledooper's suggestion to use sep=', ' in the print!

Saved 5 bytes thanks to ovs' suggestion to use bytes objects.

New Answer

print(*b"
","miss a few",*b"cd",sep=", ")

The b"
" is equivalent to b"\1\2".

Try it online!

Old Answer:

print(f"{4-3}, {6-4}, miss a few, {33*3}, {4*5*5}")

Try it online!

Vim, 37 keystrokes

i3, 3, miss a few, 8, <esc>|<C-x><C-x>l.l<C-a>ylp$pp<C-a>

TeX, 41 bytes

In TeX you can print the ASCII-value of a character using \number`CHAR.

\def~{\number`}~^A, ~^B, miss a few, ~c, ~d

Replace ^A and ^B in the above code with SOH and STX.

Using just the printable range of ASCII, the following works as well without modifications (4 bytes more):

\def~{\number`}~^^A, ~^^B, miss a few, ~c, ~d

Complete TeX file:

\def~{\number`}~, ~, miss a few, ~c, ~d
\bye

^{To copy the code above (the complete file), hit on the edit-button, so that you can copy the unprintable ASCII-letters, which are displayed in the editor.}

05AB1E, 16 bytes

I don't think this would be shorter with 0, 1, 2 and 9.

XY…š§€…†Óт<т„, ý

Try it online!

X                # push 1
 Y               # push 2
  …š§€…†Ó        # push compressed dictionary string "miss a few"
         т<      # push 100-1
           т     # push 100
            „,   # push string ", "
               ý # join the stack with this string

Zsh, 47 bytes

Not very creative. Oh well.

echo $[3/3], $[6/3], miss a few, $[##c], $[##d]

Try it online!

Whispers v3, 124..120 114 bytes

>> Each 5 57
>> Then 7 3 6 3 4 3 54 3 8
> ", "
> "miss a few"
>> "L"
>> #3
>> #5
>> 53*6
>> #4
>> 8-7
>> Output 56

Try it online!

Explanation:

I use the fact that each line with the line number \$x\$ can be called with \$x + k \cdot n \$, where \$k \in \mathbb{N} \$ and \$n\$ is total number of lines of code. Example from code with 11 lines: the reference to line 1 can be replaced by 12, 23, 34, and so on.

Try the translated version online!

Line by line:

As always in Whispers, we run the last line first:

>> Output 56

This line outputs the result from line 56. Since the code does not have 56 lines it actually outputs line \$ 56 \mod 11 = 1\$:

>> Each 5 57

Applying the same trick again we can replace this line with:

>> Each 5 2

In line 5 we can expect a function and in line 2 an array. The function will be applied on each element of the array and it replaces the element with the result. Let us first look at line 5:

>> "L"

L is an argument from the Each statement in line 1. L is converted to a string.

Now line 2:

>> Then 7 3 6 3 4 3 54 3 8

Then creates an array with the arguments as its elements. If we would print this line we get the following array:

[1, ', ', 2, ', ', 'miss a few', ', ', 99, ', ', 100]

Explanation of the arguments:

Line 3 and 4: simple strings

Line 6: Returns the length of line 3, so we get 2

Line 7: Returns the length of line 5, so we get a 1

Line 8: Returns the result of line 53 (actually 9) raised to the power of the result of line 6, so we get 100

Line 54 (actually 10): Subtracts result of line 8 with the result of line 7, so we get a 99

PHP, 48 bytes

<?=join(', ',[!'',~-3,'miss a few',33*3,45^73]);

Try it online!

Ruby, 42 bytes

$><<[3/3,6/3,'miss a few',33*3,5*5*4]*", "

Try it online!

-4 from Jonah. Borrows from Arnauld's answer.

Ruby, ⁴⁶ 44 bytes

$><<"
cd".bytes.insert(2,"miss a few")*", "

Try it online!

The three unprintables for codepoints 1, 2 and 3 are in the string containing "cd".

-2 from manatwork. Other version with same bytecount

HTML + CSS, 118 bytes

<c>, <c>, miss a few, <d><d><c>, <c><style>c:before{content:counter(a);counter-increment:a}d{counter-increment:a 48}

Pyth, 29 bytes

j+\,d[=hZhZ"miss a few"t=*TTT

Try it online!

Python 3.8 translation:

Z=0
T=10

print(", ".join(map(str,[(Z:=Z+1),Z+1,"miss a few",(T:=T*T)-1,T])))

Standard ML (MLton), 65 bytes

print(String.map(fn c=>chr(ord c-3))"4/#5/#plvv#d#ihz/#<</#433");

Try it online!

"4/#5/#plvv#d#ihz/#<</#433" is "1, 2, miss a few, 99, 100" shifted by adding three to each ASCII code. String.map(fn c=>chr(ord c-3)) reverses this shifting by substracting three of each char's ordinal number.

Jelly, 22 bytes

³ḊĖ.ịṚj“¤mɗʂɠẉ⁾»W¤j⁾, 

A full program that prints the result.

Try it online!

How?

³ḊĖ.ịṚj“¤mɗʂɠẉ⁾»W¤j⁾,  - Link: no arguments
³                      - 100
 Ḋ                     - dequeue -> [2,3,...,99,100]
  Ė                    - enumerate -> [[1,2],[2,3],...,[98,99],[99,100]]
   .                   - 0.5
    ị                  - index into -> [[99,100],[1,2]]
     Ṛ                 - reverse
                 ¤     - nilad followed by links as a nilad:
       “¤mɗʂɠẉ⁾»       -   dictionary compression of "miss a few"
                W      -   wrap in a list  -> ["miss a few"]
      j                - join              -> [1,2,"miss a few",99,100]
                   ⁾,  - ", "
                  j    - join              -> [1,", ",2,", ","miss a few",", ",99,", ",100]
                       - implicit print    -> `1, 2, miss a few, 99, 100`

R, 43 41 bytes

cat(T+F:T,"miss a few",3*33+F:T,sep=", ")

Try it online!

T and F are equal to TRUE and FALSE. They get coerced to integers 1 and 0 by the operators + and :.

brainfuck, 146 bytes

-[----->+>+>+>++>++<<<<<]>>>--.-----.------------.<-.------.>.>+++++++.>+++.<++++++..<.>>--------.<<.>>+++++.-.<++++.<<.>.<<++++++..>.>.<+++++.-..

Try it online!

A fairly straightforward 255/5=51 loop giving ASCII 51 and ASCII 102 as 333ff in 5 consecutive cells, followed by hunt and peck strategy.

Python 2, 59 bytes

print ', '.join(map(str,(4-3,5-3,'miss a few',33*3,5*5*4)))

Try it online!

Thanks to @arnauld for js answer

R, 52 bytes

cat(chartr("@-K","/-:","B, C, miss a few, JJ, BAA"))

Try it online!

Encodes numbers as uppercase letters and then applies character range translation. Since we can't directly type 0-9, we expand the range by 1 character in both directions resulting in /-:.

Bash + coreutils, 40

tr 3-? /-:<<<"5, 6, miss a few, ==, 544"

Try it online!

Jelly, 23 bytes

3Ṗ,³ṭ’¤j“¤mɗʂɠẉ⁾»W¤j⁾, 

Try it online!

How it works

One thing to note: Jelly has the Ṙ atom, which prints a string Jelly representation of it's argument. Lists in Jelly do not use [ and ] as open/close markers, instead they consist of comma-separated values:

1,2,3,4,5Ṙ

prints 1,2,3,4,5, unfortunately, without spaces.

3Ṗ,³ṭ’¤j“¤mɗʂɠẉ⁾»W¤j⁾,  - Main link. Takes no arguments
3                       - Yield 3;
 Ṗ                      - Pop from 3, yielding [1, 2]
      ¤                 - Group the previous links into a nilad:
   ³                    -   100
     ’                  -   Decrement; 99
    ṭ                   -   Tack; [99, 100]
  ,                     - Pair; [[1, 2], [99, 100]]
                  ¤     - Group the previous links into a nilad:
        “¤mɗʂɠẉ⁾»       -   The compressed string "miss a few"
                 W      -   Wrap; ["miss a few"]
       j                - Join; [1, 2, ["miss a few"], 99, 100]
                   j⁾,  - Join with ", "; "1, 2, miss a few, 99, 100"

Jelly, 23 bytes

“×ƤṪ5Ġpṃ:Ƒ5ṫȷØJkxXp⁺¥ġ»

Try it online!

Same length as caird's answer, much more boring.

Husk, 31 bytes

mȯ←←←"4/#5/#plvv#d#ihz/#<</#433

Try it online!

Explanation:

m                                 Map the function...
 ȯ←←←                             Decrement character 3 times
      "4/#5/#plvv#d#ihz/#<</#433  desired string "ASCII-shifted" by 3

PowerShell (ASCII answer), 128 bytes

$k=3-3;((@(46+3),@(53-3),@('miss a few'),@(57,57),@((46+3),48,48))|%{$k++;if($k-eq3){$_}else{($_|%{[char]$_})-join''}})-join', '

Try it online!

PowerShell (Base64 one), 135 bytes

[Text.Encoding]::Unicode.GetString([Convert]::FromBase64string('MQAsACAAMgAsACAAbQBpAHMAcwAgAGEAIABmAGUAdwAsACAAOQA5ACwAIAAxADAAMAA='))

Try it online!

Julia, 41 bytes

()->join([b"
";"miss a few";b"cd"],", ")

Try it online!

the first string is "\1\2" which is allowed if I understand the challenge correctly

b"str" converts the string in an array of the value of each character so b"\1\2" => [1,2] and b"cd" => [99,100]

Retina 0.8.2, 32 bytes

B, C, miss a few, JJ, BAA
T`L`d

Try it online! Explantion:

B, C, miss a few, JJ, BAA

Replace the empty input with the literal string.

T`L`d

Transliterate upper case letters to digits.

JavaScript (ES6), 43 bytes

Browser only.

_=>atob`MSwgMiwgbWlzcyBhIGZldywgOTksIDEwMA`

Try it online! (with a polyfill)

JavaScript (ES6), 45 bytes

_=>[4-3,5-3,'miss a few',33*3,5*5*4].join`, `

Try it online!