Compare

Claude Haiku 4.5 vs GPT-5 Mini

Anthropic vs OpenAI. Specs, benchmarks, and real per-task cost — all in one page.

Verdict

Claude Haiku 4.5 leads on LMSYS ELO (1305 vs 1288). GPT-5 Mini is ~1.8× cheaper on a 3:1 input:output blend.

Anthropic

Fast, cheap, and smart enough for most routing and extraction tasks. Great base model for subagents.

Sub-second latency10x cheaper than OpusStrong structured output

OpenAI

GPT-5’s smaller, cheaper sibling. Optimized for chat and lightweight agent tasks.

Very low latencyCheap enough for mass consumer apps

Pricing

Input / 1M$0.80$0.50

Output / 1M$4.00$2.00

Context200K256K

Max output16K16K

LicenseProprietaryProprietary

Released2025-10-012026-01-21

Benchmarks

LMSYS ELO1305.01288.0

MMLU Pro81.478.5

HumanEval87.085.5

SWE-Bench49.2—

MATH76.879.8

IFEval87.385.4

Per-task cost

Summarize a 1-hour meeting transcript$0.014$0.0085

Review a 500-line pull request$0.014$0.0080

Answer a customer support ticket$0.0056$0.0031

Extract structured data from a resume$0.0080$0.0045

Debug a stack trace with context$0.013$0.0070

Per-call cost using published token counts for each task. Real-world prompts vary.

Add up to four models, tweak tasks live.