Claude Haiku 4.5
#104 of 274 scored.
AnthropicOct 15, 2025Closed
Capabilities Index
Range 139.4–144.1
142.4
Context
Takes text, image, file
200K
Input
Per 1M tokens
$1.00
Output
Per 1M tokens
$5.00
Benchmarks
How it scores.
- GPQA Diamond61.6%
99th of 186
- ARC-AGI-24.0%
57th of 79
- SimpleQA Verified13.2%
70th of 77
- Mock AIME66.6%
98th of 176
- Terminal-Bench35.5%
23rd of 35
Everything else Epoch has for it
- MATH level 596.4%
- DTBench56.0%
- ARC-AGI47.7%
- DeepResearch Bench45.5%
- WeirdML45.4%
- LMCA36.4%
- Balrog31.2%
- ExploitBench13.7%
- FrontierMath-2025-02-28-Private10.4%
- FrontierMath-Tier-4-2025-07-01-Private3.5%
- Chess Puzzles3.2%
On the index
Around it.
Sources