What's a Good Wordle Score? A Solver Benchmark
A four is a normal Wordle result, and an average close to four is a solid score. This site's solver averages 3.52 guesses, but software is not a fair target for a person. It can inspect a complete list and compare thousands of moves without getting tired.
A benchmark is a fixed test used for comparison. It is useful because every answer gets the same treatment. It does not tell you what your score ought to be.
How we tested this
The benchmark applies the same setup to 2,374 answers. It opens with ROATE, considers a limited Wordle guess list, then chooses the word expected to leave the fewest answers. That makes the answers comparable with one another. It is not a proven minimum and it is not a like-for-like comparison with a person playing from memory.
Read the result in context
Of the 2,374 tested answers, 60 finish in two guesses, 1,099 in three, 1,137 in four and 78 in five. The longest route uses 5 guesses.
48.8% finish by turn three and 96.7% finish by turn four. Even this systematic solver needs four guesses for 1,137 answers and a fifth for 78.
| Guesses | Answers | % of answers |
|---|---|---|
| 2 | 60 | 2.5 |
| 3 | 1,099 | 46.3 |
| 4 | 1,137 | 47.9 |
| 5 | 78 | 3.3 |
Work out your own average
Add the guesses from a group of games, then divide by the number of games. For example, scores of 3, 4, 4 and 5 add up to 16. Divide 16 by four games and the average is 4.0.
- Use enough games. One lucky two or one painful six can distort a tiny sample. A longer run says more about your usual play.
- Compare like with like. Hard Mode, outside help and skipped games change the conditions. Keep those in mind before comparing averages.
- Look for a useful pattern. If many games stall on turn two, work on narrowing the shortlist. If similar words cause sixes, practise trap-family probes.
Improve the decisions, not just the number
Early moves have the greatest effect on your average. A balanced opener usually leaves a smaller shortlist. A thoughtful second guess tests the letters that separate the survivors. When those survivors form a trap family, a splitter can be worth more than an immediate answer attempt.
The Opener Lab compares first words using the same first-turn data. Treat the 3.52 result as a consistent software baseline, not a universal definition of a good score.
Explore one row of the benchmark
Choose a score to see how often the fixed solver recorded it. This describes the benchmark, not the quality of a human player.
Without the selector, every score is summarised below.
2 guesses
The solver used 2 guesses for 60 answers, or 2.5% of this test.
3 guesses
The solver used 3 guesses for 1,099 answers, or 46.3% of this test.
4 guesses
The solver used 4 guesses for 1,137 answers, or 47.9% of this test.
5 guesses
The solver used 5 guesses for 78 answers, or 3.3% of this test.
Compare your board with the solver’s ranked moves.
Open the Solver