Anthropic · proprietary

Claude Haiku 4.5

Claude Haiku 4.5 by Anthropic appears in 1 source with Reasoning at 35.4. Best read for Code quality, LLM, Low cost.

CreatorAnthropic

Release date2025-10-15

Knowledge cutoffNot published

Context200K tokens

Input price$1/M tokens

Output price$5/M tokens

Modalitytext + vision

CountryUS

Metrics

All source-backed metrics

Open in leaderboard

LLM Stats Rank50ranking · source_rank

Reasoning35.4reasoning · index_reasoning

Math20.89math · index_math

Coding25.09coding · index_code

Writing25.23writing · index_communication

Vision16.63multimodal · index_vision

Tool calling21.32tool_calling · index_tool_calling

Long context15.04long_context · index_long_context

Healthcare-0.1domain · index_healthcare

GPQA73 %reasoning · gpqa_score

AIME 202580.7 %math · aime_2025_score

SWE-bench Verified73.3 %coding · swe_bench_verified_score

Code Arena963.43 %coding · coding_arena_score

MMMLU83 %reasoning · mmmlu_score

OSWorld50.7 %agent · osworld_score

Terminal Bench41 %coding · terminal_bench_score

Context200,000 tokenscontext · context

Speed222.37 c/sperformance · throughput

Input price1 $/Mpricing · input_price

Output price5 $/Mpricing · output_price

Evidence

Citations and source overlap

LLM StatsPublic full LLM leaderboard advanced explorer snapshot, including ranking, benchmark indexes, pricing, speed, context, license, and model metadata.Retrieved 2026-06-01

FAQ

How should I read this profile?

Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.

Is Claude Haiku 4.5 verified by EvalKit?

No. EvalKit currently shows 0 verified rows until real run evidence exists.

Why can metrics disagree?

Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.