Claude 4.5 Haiku
Anthropic · 2 thinking configurations
Intelligence16.9
Input / 1M$1.00
Output / 1M$5.00
Context200K
Tariff: Anthropic · Source: artificial-analysis · USD per million tokens
Thinking configuration
What does more thinking buy you?
Compare this model’s measured thinking configurations.
Scroll the chart ↔ · tap a point for details
Efficient frontier: best measured score at each cost1 measured configurations · 1 without both values
Selected: thinking thinking · Click a point to switch configuration
Claude Haiku 4.5 is Anthropic’s fastest and most efficient model, delivering near-frontier intelligence at a fraction of the cost and latency of larger Claude models. Matching Claude Sonnet 4’s performance...
Intelligence IndexArtificial Analysis composite index, not a percentage. Different versions of the index are not directly comparable.
16.9 pointsGPQA DiamondGraduate-level scientific questions. Measures specialist reasoning rather than web-search quality.
67.2 %Humanity’s Last ExamDifficult questions across many disciplines, compared under the same evaluation setup.
10.4 %SciCodeCode generation for scientific research problems.
42.2 %Terminal-Bench 2.1Agent tasks in a terminal environment. Results also depend on the evaluation harness.
44.2 %τ²-BenchTool use and interaction in assistance scenarios. Does not replace an evaluation of your integration.
54.7 %IFBenchVerifiable instruction following. A proxy for precision rather than a benchmark of creative style.
54.3 %MMMU-ProMultimodal understanding of academic problems that include images.
58.6 %Omniscience accuracyFactual accuracy in Artificial Analysis's Omniscience evaluation.
18.0 %Specifications from the catalog
Inputtext, image, file
Outputtext
Tool callingYes
Structured outputYes
Speed89.3 tok/s
First token15.75 s
Maximum output64K
Catalog listing date2025-10-15