Claude Haiku 4.5 vs Claude Sonnet 4.5
Wins 1 of 7 areas
Long documents
Wins 5 of 7 areas
Coding · Agents · Reasoning · Images and charts · Following instructions
Claude Sonnet 4.5 is the stronger all-rounder.Claude Haiku 4.5 is cheaper and better at long documents.
Scores updated · 45 tests both models report · How we compare
Where each one wins
Tests won in each of the seven 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.
- CodingWriting and fixing software06Claude Sonnet 4.56 of 7 tests · 1 tie
- ReasoningHard problems that need careful thinking04Claude Sonnet 4.54 of 4 tests
- AgentsCarrying out multi-step tasks on its own03Claude Sonnet 4.53 of 4 tests · 1 tie
- Following instructionsDoing exactly what it is asked02Claude Sonnet 4.52 of 2 tests
- Images and chartsUnderstanding pictures, charts and video01Claude Sonnet 4.51 of 1 test
- Long documentsFinding answers in very long texts10Claude Haiku 4.51 of 1 test
- FactsGetting facts right instead of making them up22Even2 each
Images and charts and long documents rest on a single test each.
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.
Claude Haiku 4.5 costs 67% less for the same work.
The biggest differences
The tests each model wins by the widest margin, up to three each. Scores are out of 100.
Where Claude Sonnet 4.5 pulls ahead
- Short factual questions, answered correctlySimpleQA Verified+17.5points ahead
- Graduate-level biology, physics and chemistry questionsGPQA Diamond+16.2points ahead
- Customer-service tasks in a simulated bankτ-Bench V3 · Banking+15.2points ahead
Where Claude Haiku 4.5 pulls ahead
- Avoids making up answers it doesn't knowAA-Omniscience · Non-hallucination+21.8points ahead
- Reasons across sets of long documentsAA-LCR+2points ahead
Every test, side by side
All 45 tests both models report. The winning score is in its model's colour; marks a score checked independently.
CodingClaude Sonnet 4.5
- Terminal-Bench 2.1Claude Sonnet 4.5 by 11.644.255.8+11.6
- Terminal-Bench HardClaude Sonnet 4.5 by 8.327.335.6+8.3
- LMArena · WebDevClaude Sonnet 4.5 by 64 rating points13291393+64 rating
- SWE-bench ProClaude Sonnet 4.5 by 4.139.543.6+4.1
- SWE-bench VerifiedClaude Sonnet 4.5 by 3.973.377.2+3.9
- SciCodeClaude Sonnet 4.5 by 3.542.245.7+3.5
- SWE-bench Multilingualtie67.467tie
AgentsClaude Sonnet 4.5
- τ-Bench V3 · BankingClaude Sonnet 4.5 by 15.29.324.5+15.2
- OSWorld-VerifiedClaude Sonnet 4.5 by 10.750.761.4+10.7
- GDPValClaude Sonnet 4.5 by 8.711.820.5+8.7
- Terminal-Bench 4.0tie00tie
ReasoningClaude Sonnet 4.5
- GPQA DiamondClaude Sonnet 4.5 by 16.267.283.4+16.2
- ARC-AGI-2Claude Sonnet 4.5 by 9.6413.6+9.6
- Humanity's Last ExamClaude Sonnet 4.5 by 7.410.417.8+7.4
- CritPtClaude Sonnet 4.5 by 1.101.1+1.1
FactsEven
- AA-Omniscience · Non-hallucinationClaude Haiku 4.5 by 21.872.750.9+21.8
- SimpleQA VerifiedClaude Sonnet 4.5 by 17.513.230.7+17.5
- AA-Omniscience · AccuracyClaude Sonnet 4.5 by 14.91832.9+14.9
- Vectara HHEM hallucination ratelower is betterClaude Haiku 4.5 by 2.29.812+2.2
Images and chartsClaude Sonnet 4.5
- MMMU-ProClaude Sonnet 4.5 by 10.158.668.7+10.1
Following instructionsClaude Sonnet 4.5
- Multi-ChallengeClaude Sonnet 4.5 by 4.850.555.3+4.8
- IFBenchClaude Sonnet 4.5 by 354.357.3+3
Other results22 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)Claude Sonnet 4.5 by 23.454.778.1+23.4
- ARC-AGI-1Claude Sonnet 4.5 by 1647.763.7+16
- BrowseComp-Plus (with tool result clearing + memory)Claude Sonnet 4.5 by 12.554.767.2+12.5
- Terminal-BenchClaude Sonnet 4.5 by 94150+9
- Artificial Analysis Coding IndexClaude Sonnet 4.5 by 8.243.952.1+8.2
- BrowseComp-Plus (with tool result clearing)Claude Sonnet 4.5 by 7.952.560.4+7.9
- AA Agentic IndexClaude Sonnet 4.5 by 7.210.317.5+7.2
- AIME 2025Claude Sonnet 4.5 by 6.380.787+6.3
- bbqClaude Sonnet 4.5 by 6.192.898.9+6.1
- MMMLUClaude Sonnet 4.5 by 6.18389.1+6.1
- AA-OmniscienceClaude Sonnet 4.5 by 4.3-4.4-0.1+4.3
- HarmBenchClaude Haiku 4.5 by 3.995.992+3.9
- vectara_answer_rateClaude Haiku 4.5 by 3.999.595.6+3.9
- AA IntelligenceClaude Sonnet 4.5 by 3.816.920.7+3.8
- AIR-Bench 2024Claude Haiku 4.5 by 3.493.289.8+3.4
- XSTestClaude Sonnet 4.5 by 3.293.296.4+3.2
- τ²-Bench (Retail)Claude Sonnet 4.5 by 383.286.2+3
- vectara_factual_consistencyClaude Haiku 4.5 by 2.290.288+2.2
- frontiermath_tier_4_v1Claude Sonnet 4.5 by 2.12.14.2+2.1
- anthropic_red_teamClaude Sonnet 4.5 by 1.396.998.2+1.3
- vectara_avg_summary_lengthClaude Sonnet 4.5 by 12.7 rating points115.1127.8+12.7 rating
- simple_safety_testsClaude Sonnet 4.5 by 1.298.8100+1.2
Questions people ask
Which is better, Claude Haiku 4.5 or Claude Sonnet 4.5?
Claude Sonnet 4.5 wins five of the seven areas where both have results: coding, agents, reasoning, images and charts and following instructions. Claude Haiku 4.5 wins long documents, and costs 67% less. They are level on facts.
Which is better for coding?
Claude Sonnet 4.5. It wins 6 of the 7 coding tests both models report; Claude Haiku 4.5 wins none, and 1 is a tie.
Which is cheaper?
Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens; Claude Sonnet 4.5 costs $3.00 and $15.00. That makes Claude Haiku 4.5 about 67% cheaper for the same work.
How do you compare the two?
We use the 45 benchmark tests both models have published scores on. The verdict counts the 23 tests in the eight capability areas, and a gap under one point (ten on rating-style scales) counts as a tie. The other 22 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.