Claude Haiku 4.5 vs Claude Sonnet 4.6
Side-by-side benchmark comparison across coding, math, reasoning, speed, and pricing.
Claude Sonnet 4.6 by Anthropic wins on 13 of 18 benchmarks against Claude Haiku 4.5 by Anthropic, which leads on 5. This head-to-head comparison covers coding, math, reasoning, speed, and pricing metrics from our benchmark data.
Category-by-Category Breakdown
General Intelligence: In general intelligence, Claude Sonnet 4.6 scores 1395 on Chatbot Arena ELO compared to Claude Haiku 4.5's 1220, while Claude Sonnet 4.6 scores 84.0% on MMLU-Pro compared to Claude Haiku 4.5's 65.0%, while Claude Sonnet 4.6 scores 30.5 on AA Intelligence Index compared to Claude Haiku 4.5's 17.6.
Coding: In coding, Claude Sonnet 4.6 scores 91.0% on HumanEval+ compared to Claude Haiku 4.5's 78.0%, while Claude Sonnet 4.6 scores 79.6% on SWE-bench Verified compared to Claude Haiku 4.5's 30.0%, while Claude Sonnet 4.6 scores 65.0% on LiveCodeBench compared to Claude Haiku 4.5's 35.0%, while Claude Sonnet 4.6 scores 63 on AA Coding Index compared to Claude Haiku 4.5's 43.9.
Math: In math, Claude Sonnet 4.6 scores 88.0% on MATH compared to Claude Haiku 4.5's 70.0%, while Claude Sonnet 4.6 scores 96.0% on GSM8K compared to Claude Haiku 4.5's 88.0%.
Reasoning: In reasoning, Claude Sonnet 4.6 scores 74.1% on GPQA Diamond compared to Claude Haiku 4.5's 40.0%, while Claude Sonnet 4.6 scores 48.0% on ARC-AGI compared to Claude Haiku 4.5's 18.0%, while Claude Sonnet 4.6 scores 33.1 on AA Agentic Index compared to Claude Haiku 4.5's 10.3.
Context: In context, Claude Sonnet 4.6 scores 1.0M on Context Length compared to Claude Haiku 4.5's 200K.
Pricing Comparison
Claude Haiku 4.5 costs $1.0/1M input tokens and $5.0/1M output tokens, while Claude Sonnet 4.6 costs $3.0/1M input and $15.0/1M output. Claude Haiku 4.5 is the more affordable option for API usage.
Speed Comparison
Claude Haiku 4.5 generates output at 180 tok/s compared to Claude Sonnet 4.6's 57 tok/s, and the time to first token is 120 ms for Claude Haiku 4.5 versus 790 ms for Claude Sonnet 4.6. Claude Haiku 4.5 delivers faster throughput.
Verdict
For developers prioritizing affordability and speed, Claude Haiku 4.5 has the edge. For those who value coding and general intelligence and math, Claude Sonnet 4.6 is the stronger choice.
View Individual Model Pages
Claude Haiku 4.5 vs Claude Sonnet 4.6 — FAQ
Which is better, Claude Haiku 4.5 or Claude Sonnet 4.6?
Claude Sonnet 4.6 wins on more benchmarks overall (13 vs 5). However, the best choice depends on your specific needs — each model excels in different areas.
How does Claude Haiku 4.5 compare to Claude Sonnet 4.6 for coding?
Claude Sonnet 4.6 is better for coding, scoring 79.6% on SWE-bench Verified compared to 30.0%. SWE-bench tests real-world software engineering by resolving actual GitHub issues.
Is Claude Haiku 4.5 cheaper than Claude Sonnet 4.6?
Yes, Claude Haiku 4.5 is cheaper. Claude Haiku 4.5 costs $1.0/1M input and $5.0/1M output tokens. Claude Sonnet 4.6 costs $3.0/1M input and $15.0/1M output tokens.
Which is faster, Claude Haiku 4.5 or Claude Sonnet 4.6?
Claude Haiku 4.5 is faster, generating output at 180 tok/s compared to 57 tok/s. Faster output speed means shorter wait times for API responses.
What benchmarks does the Claude Haiku 4.5 vs Claude Sonnet 4.6 comparison cover?
This comparison covers 18 benchmarks including Chatbot Arena ELO, MMLU-Pro, HumanEval+, MATH, GPQA Diamond, SWE-bench Verified, Output Speed, LiveCodeBench, and more. Metrics span general intelligence, coding, math, reasoning, speed, and cost categories.