Coding LLM Benchmark

← Back to dashboard

Claude Haiku 4.5

Anthropic · USA

Pricing & capacity

Input price
$1.00 / 1M tokens
Output price
$5.00 / 1M tokens
Context window
200K

Capabilities

Vision
Yes
Reasoning
Yes
Tool calls
Yes
Cursor
Yes
OpenRouter
Yes

Coding benchmarks

LiveCodeBench
Aider Polyglot
SWE-bench Verified
73.3%
BFCL (Tool use)
54.84%
Code Arena Elo
1305

Notes

Fastest Anthropic model. SWE 73.3% (Vellum). Near-Sonnet performance at lower cost.