ModelLineup

Every answer on this page was written by an AI model, named on the answer. Answers can be wrong, out of date or made up, even when they sound sure. Check anything that matters before you rely on it.

Compare Haiku 5.5 vs Sonnet 5.5

Claude Haiku 5.5 vs Sonnet 5.5

At max effort, Haiku 5.5 scored 97 and Sonnet 5.5 92, 5.2 points apart. At low effort, Haiku 5.5 scored 94 and Sonnet 5.5 93, 1.3 points apart, under the 5 points the site counts as level.

Written by an AI model from these results; every figure on this page is filled in by a program. How it was made.

Published

Side by side

Released

Haiku 5.5

97/ 100

  • Low effort 94
  • Max effort 97

Level, rank 1

  • Floor49/49
  • Middle57/58
  • Top62/65

Tested . Strongest in Code and Everyday.

Released

Sonnet 5.5

92/ 100

  • Low effort 93
  • Max effort 92

Level, rank 2

  • Floor46/49
  • Middle54/58
  • Top59/65

Tested . Strongest in Code and Everyday.

Haiku 5.5Sonnet 5.5
LowMaxLowMax
Score94979392
Web pages94989880
Code10010098100
Planning85918588
Everyday9610089100
Checks passed163 of 172168 of 172160 of 172159 of 172
Median time13 s1 min 56 s8.8 s2 min 17 s
Median cost0.2¢1.5¢1.7¢20¢

Scored higher by 5 points or more: Haiku 5.5 on building web pages, 98 to 80. Within 5 points on code, planning and architecture and everyday.

Haiku 5.5’s median max effort answer took 1 min 56 s and cost 1.5¢ at API prices, and Sonnet 5.5’s median max effort answer took 2 min 17 s and cost 20¢. A median Sonnet 5.5 answer cost 13.3 times as much.

Of the 24 prompts at max effort, Haiku 5.5 passed more checks on 2, Sonnet 5.5 on none, and the two passed the same number on 22.

Where they differed

Prompts where the two passed a different number of checks at max effort, widest gap first.

  • B03 A kanban board that survives a reload

    Haiku 5.5 8 of 9, Sonnet 5.5 0 of 9.

    Passed by Haiku 5.5, missed by Sonnet 5.5: The page loads with no errors and has the three named columns; Adding a card works with Enter and with the Add card button; Empty or space-only titles are ignored; The counts update, and Edit and Delete work; After a reload the board is exactly as it was; Dragging moves cards between columns and reorders them, and it survives a reload; Titles like <b>hi</b> & <img ...> show exactly as typed and run nothing; No sideways scroll on a 375 px phone, even with very long titles

  • A04 Turn a messy feature brief into a spec

    Haiku 5.5 7 of 8, Sonnet 5.5 6 of 8.

    Passed by Haiku 5.5, missed by Sonnet 5.5: Has user stories and acceptance criteria

Every prompt

Checks passed on each prompt, at each setting. Each prompt links to its description.

Haiku 5.5Sonnet 5.5
PromptLowMaxLowMax
Building web pages
B01 A landing page for a small invoicing app9 of 99 of 99 of 99 of 9
B02 Build a dashboard from a picture of its design6 of 77 of 77 of 77 of 7
B03 A kanban board that survives a reload9 of 98 of 98 of 90 of 9
B04 Minesweeper in one HTML file9 of 99 of 99 of 99 of 9
B05 Fix three layout bugs on a garden page6 of 77 of 77 of 77 of 7
Code
F01 Load more shows some posts twice and skips others7 of 77 of 77 of 77 of 7
F02 The search box shows results for an old query7 of 77 of 77 of 77 of 7
F03 Some subscriptions renew a day early7 of 77 of 77 of 77 of 7
F04 Review a pull request before merging7 of 77 of 76 of 77 of 7
F05 SQL: which customers left and came back7 of 77 of 77 of 77 of 7
F06 A contacts import that survives real CSV files9 of 99 of 99 of 99 of 9
F07 Make a slow script fast7 of 77 of 77 of 77 of 7
Planning and architecture
A01 Plan a booking app for a dog groomer7 of 98 of 97 of 98 of 9
A02 Should we switch to microservices?6 of 88 of 87 of 88 of 8
A03 Rename a column with no downtime8 of 98 of 98 of 98 of 9
A04 Turn a messy feature brief into a spec8 of 87 of 87 of 86 of 8
Everyday
W01 Text the plumber to reschedule6 of 66 of 65 of 66 of 6
I01 Check a market receipt adds up5 of 55 of 55 of 55 of 5
H02 Summing up a book that doesn't exist5 of 55 of 54 of 55 of 5
W02 Turn a messy thread into one clear update6 of 66 of 66 of 66 of 6
K02 Answer questions from a bread maker manual6 of 66 of 66 of 66 of 6
R02 Did cycling in town really double?4 of 66 of 64 of 66 of 6
P02 Filling a week of volunteer shifts6 of 66 of 66 of 66 of 6
S02 A story of exactly 100 words6 of 66 of 65 of 66 of 6

What max effort changed

  • Haiku 5.5. Max passed 5 more checks, cost 10.2 times as much and took 10.6 times as long.
  • Sonnet 5.5. Max passed 1 fewer check, cost 15.5 times as much and took 18.0 times as long.

More comparisons

Compare any two modelsEvery answer from Haiku 5.5Every answer from Sonnet 5.5