Claude 3.5 Haiku vs HyperCLOVA X SEED Think (32B)
Wins 2 of 5 areas
Long documents · Following instructions
Wins 2 of 5 areas
Coding · Reasoning
The two are evenly matched.Claude 3.5 Haiku is better at long documents and following instructions; HyperCLOVA X SEED Think (32B) at reasoning and coding.
Scores updated · 12 tests both models report · How we compare
Where each one wins
Tests won in each of the five 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.
- ReasoningHard problems that need careful thinking02HyperCLOVA X SEED Think (32B)2 of 3 tests · 1 tie
- CodingWriting and fixing software02HyperCLOVA X SEED Think (32B)2 of 2 tests
- Long documentsFinding answers in very long texts10Claude 3.5 Haiku1 of 1 test
- Following instructionsDoing exactly what it is asked10Claude 3.5 Haiku1 of 1 test
- FactsGetting facts right instead of making them up11Even1 each
Long documents and following instructions rest on a single test each.
The biggest differences
The three tests each model wins by the widest margin. Scores are out of 100.
Where HyperCLOVA X SEED Think (32B) pulls ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+20.7points ahead
- Hard command-line tasks in a real terminalTerminal-Bench Hard+9.8points ahead
- Answers hard knowledge questions correctlyAA-Omniscience · Accuracy+2.2points ahead
Where Claude 3.5 Haiku pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+40.1points ahead
- Reasons across sets of long documentsAA-LCR+13.6points ahead
- Follows unfamiliar, precisely checkable instructionsIFBench+4.9points ahead
Every test, side by side
All 12 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingHyperCLOVA X SEED Think (32B)
- Terminal-Bench HardHyperCLOVA X SEED Think (32B) by 9.82.312.1+9.8
- SciCodeHyperCLOVA X SEED Think (32B) by 127.428.4+1
ReasoningHyperCLOVA X SEED Think (32B)
- GPQA DiamondHyperCLOVA X SEED Think (32B) by 20.740.861.5+20.7
- Humanity's Last ExamHyperCLOVA X SEED Think (32B) by 1.93.65.5+1.9
- CritPttie00tie
FactsEven
- AA-Omniscience · Non-hallucinationClaude 3.5 Haiku by 40.158.918.8+40.1
- AA-Omniscience · AccuracyHyperCLOVA X SEED Think (32B) by 2.213.215.4+2.2
Long documentsClaude 3.5 Haiku
- AA-LCRClaude 3.5 Haiku by 13.627.313.7+13.6
Following instructionsClaude 3.5 Haiku
- IFBenchClaude 3.5 Haiku by 4.942.837.9+4.9
Other results3 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.
- τ²-Bench Telecom (AA run)HyperCLOVA X SEED Think (32B) by 62.824.687.4+62.8
- AA-OmniscienceClaude 3.5 Haiku by 30.8-22.5-53.3+30.8
- AA IntelligenceHyperCLOVA X SEED Think (32B) by 2.58.911.4+2.5
Questions people ask
Which is better, Claude 3.5 Haiku or HyperCLOVA X SEED Think (32B)?
Claude 3.5 Haiku and HyperCLOVA X SEED Think (32B) each win two of the five areas where both have results. Claude 3.5 Haiku wins long documents and following instructions; HyperCLOVA X SEED Think (32B) wins coding and reasoning. They are level on facts.
Which is better for coding?
HyperCLOVA X SEED Think (32B). It wins 2 of the 2 coding tests both models report; Claude 3.5 Haiku wins none.
How do you compare the two?
We use the 12 benchmark tests both models have published scores on. The verdict counts the 9 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 3 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.