The 5,000 most common English words
Ranked by how often they actually turn up in spoken dialogue, not by how useful somebody guessed they would be.
Every number below is counted from a corpus, and the corpus is named at the bottom.
What each step buys you
Learn the first 100 words of English and you follow 59.0% of everyday speech. The first 1000 take you to 83.8%. After that every thousand words buys less than the thousand before it, which is exactly why the order matters.
| Words known | Speech followed | |
|---|---|---|
| 10 | 23.2% | |
| 50 | 47.3% | |
| 100 | 59.0% | |
| 250 | 70.9% | |
| 500 | 78.1% | |
| 1,000 | 83.8% | |
| 2,000 | 88.6% | |
| 3,000 | 91.0% | |
| 5,000 | 93.7% | |
| 8,000 | 95.7% | |
| 10,000 | 96.5% | |
| 15,000 | 97.8% | |
| 20,000 | 98.5% |
Comfortable reading starts around 98%. Everything before that is still worth doing, but the first hundred words buy you a foothold, not comprehension.
The first 100 words of English
In order. The percentage is the share of ordinary speech you cover once you know that word and every word above it.
- 1you3.97%
- 2i7.71%
- 3the10.84%
- 4to13.2%
- 5a15.2%
- 6's17.17%
- 7it19.05%
- 8and20.51%
- 9that21.92%
- 10't23.24%
- 11of24.47%
- 12is25.49%
- 13in26.51%
- 14what27.46%
- 15we28.39%
- 16me29.28%
- 17this30.07%
- 18he30.83%
- 19for31.54%
- 20my32.23%
- 21on32.89%
- 22have33.55%
- 23your34.18%
- 24do34.79%
- 25was35.4%
- 26'm36.0%
- 27no36.61%
- 28not37.2%
- 29be37.78%
- 30are38.36%
- 31don38.93%
- 32're39.49%
- 33know40.03%
- 34can40.55%
- 35with41.08%
- 36but41.58%
- 37all42.07%
- 38so42.54%
- 39just43.0%
- 40there43.44%
- 41here43.87%
- 42they44.29%
- 43like44.7%
- 44'll45.1%
- 45get45.5%
- 46she45.89%
- 47go46.26%
- 48if46.63%
- 49right46.98%
- 50out47.33%
- 51about47.67%
- 52up48.01%
- 53at48.34%
- 54him48.68%
- 55now49.0%
- 56oh49.31%
- 57one49.63%
- 58come49.93%
- 59well50.23%
- 60her50.53%
- 61how50.82%
- 62yeah51.1%
- 63've51.37%
- 64will51.64%
- 65got51.91%
- 66want52.18%
- 67think52.44%
- 68as52.68%
- 69see52.93%
- 70did53.17%
- 71good53.41%
- 72who53.65%
- 73why53.89%
- 74from54.12%
- 75let54.36%
- 76his54.59%
- 77yes54.81%
- 78when55.02%
- 79going55.23%
- 80time55.43%
- 81an55.63%
- 82okay55.83%
- 83back56.03%
- 84look56.21%
- 85us56.4%
- 86would56.58%
- 87them56.76%
- 88where56.95%
- 89were57.13%
- 90take57.31%
- 91then57.49%
- 92had57.66%
- 93or57.84%
- 94been58.01%
- 95our58.17%
- 96gonna58.34%
- 97tell58.5%
- 98really58.66%
- 99man58.82%
- 100some58.99%
Open all 5,000 words in the codex →
In the codex you tap the ones you already know and the number moves. It stays in your browser, no account.
Where English comes from
Proto-Indo-European › Germanic › West Germanic › English
Germanic bones, Norse in the joints, and half its vocabulary borrowed from French after 1066.
It never stopped taking words.
Words English has that English never built
- petrichor PET-ri-korThe smell of rain hitting dry earth
- gloaming GLO-mingThe last light after sunset, before the dark
- faff faffBusy activity that achieves nothing
- kerfuffle ker-FUF-fulA small commotion over something not worth it
Where the numbers come from
Frequencies counted over the OpenSubtitles 2018 dialogue corpus, 725,119,374 tokens of English in total. Lists compiled by Hermit Dave, released under CC BY-SA 3.0. Subtitles skew towards conversation, which is the point: this is the language people speak, not the language people publish.