Haiku 5.5 is the stronger all-rounder.
Scores updated · 8 tests both models report · How we compare
Where each one wins
Tests won in each of the two areas where both have results. Each piece is one test, so a longer bar means more evidence; grey means the two scored within a point of each other.
Reasoning rests on a single test.
What it costs
Prices per million tokens, roughly 750,000 words. The bars show the cost of a million tokens read plus a million written.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Haiku 5.5 pulls ahead
- Very hard expert questions across many subjectsHumanity's Last Exam+16.1points ahead
- Fixes real GitHub issues in many programming languagesSWE-bench Multilingual+5.4points ahead
- Long, multi-file coding tasks in real codebasesSWE-bench Pro+1.6points ahead
Where Claude Sonnet 5 pulls ahead
No clear win on a test scored out of 100.
Every test, side by side
All 8 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingHaiku 5.5
- SWE-bench MultilingualHaiku 5.5 by 5.478.383.7+5.4
- SWE-bench ProHaiku 5.5 by 1.663.264.8+1.6
Other results5 tests, not counted
Tests outside the eight areas. They are not counted above: several are summary scores built from other tests, or the same test under another name.
- Terminal-Bench 4.0Haiku 5.5 by 26.812.439.2+26.8
- GDPval-AA 2.1Haiku 5.5 by 171 rating points14491620+171 rating
- OSWorld 2.1Haiku 5.5 by 15.45772.4+15.4
- Program BenchHaiku 5.5 by 4.777.382+4.7
- SWE-bench MultimodalHaiku 5.5 by 2.628.130.7+2.6
Questions people ask
Which is better, Claude Sonnet 5 or Haiku 5.5?
Haiku 5.5 wins all two areas where both have results: coding and reasoning. Claude Sonnet 5 wins none.
Which is better for coding?
Haiku 5.5. It wins 2 of the 2 coding tests both models report; Claude Sonnet 5 wins none.
How do you compare the two?
We use the 8 benchmark tests both models have published scores on. The verdict counts the 3 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 5 are listed but not counted, because several are summary scores or repeat a test. Each score is the one shown on the model's own page.