Skip to content
VECTOR WIREAI INTELLIGENCE
⌘K
UTC
Models741Deals30d261Policy30d65, 14 pendingSources1,013

Claude Haiku 4.5

AnthropicReleased Oct 15, 2025Current

Score basis

Claude Haiku 4.5 is capable in long context; and behind the leaders in multimodal tasks, coding, and instruction following. Too few results yet to rate reasoning, agentic tasks, safety, math, multilingual tasks, or factuality.

Capability profile

Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.

Capable1 capability
  1. Long Context−19.8%1 of 3
Limited3 capabilities
  1. Multimodal−37.1%1 of 6
  2. Coding−38.6%7 of 10
  3. Instruction Following−39.2%2 of 3
Not rated6 capabilities

Too few results yet to rate Reasoning, Agentic, Safety, Math, Multilingual or Factuality.

Price

$1.00input$5.00outputper million tokens

From Anthropic's own price page · 4 providers tracked · All prices

Costs more than 71% of 331 priced models · 3:1 input-to-output blend, log scale

Evidence

81results on57benchmarks

  • 27 independently verified
  • 34 aggregator
  • 19 vendor-reported
  • 1 cross-referenced

From 22 sources · latest Oct 8, 2026 · How verification works

API features

Tool callingStructured outputsReasoning

As listed by OpenRouter

Research

59 papers reference Claude Haiku 4.5

Claude Haiku 4.5 benchmark results

81 results on 57 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.

Long Context

CapableFull long context ranking

19.8% behind the leader1 of 3 ranked benchmarks measured

Show 1 more long context resultHide 1 long context result

Multimodal

LimitedFull multimodal ranking

37.1% behind the leader1 of 6 ranked benchmarks measured

Show 1 more multimodal resultHide 1 multimodal result

Coding

LimitedFull coding ranking

38.6% behind the leader7 of 10 ranked benchmarks measured

Show 2 more coding resultsHide 2 coding results

Instruction Following

LimitedFull instruction following ranking

39.2% behind the leader2 of 3 ranked benchmarks measured

Show 1 more instruction following resultHide 1 instruction following result

Reasoning

Not enough dataFull reasoning ranking

0 of 6 ranked benchmarks measured

Show 8 more reasoning resultsHide 8 reasoning results

Agentic

Not enough dataFull agentic ranking

0 of 7 ranked benchmarks measured

Show 4 more agentic resultsHide 4 agentic results

Factuality

Not enough dataFull factuality ranking

0 of 4 ranked benchmarks measured

Show 3 more factuality resultsHide 3 factuality results

More results

Benchmarks outside the capability baskets. They are not ranked against a leader.

Show 35 more resultsHide 35 results

Claude Haiku 4.5: common questions

Who makes Claude Haiku 4.5?

Claude Haiku 4.5 is made by Anthropic.

When was Claude Haiku 4.5 released?

Claude Haiku 4.5 was released on Oct 15, 2025, according to Artificial Analysis.

What is Claude Haiku 4.5 good at?

Claude Haiku 4.5 is capable in long context; and behind the leaders in multimodal tasks, coding, and instruction following. Too few results yet to rate reasoning, agentic tasks, safety, math, multilingual tasks, or factuality.

How much does Claude Haiku 4.5 cost?

Claude Haiku 4.5 costs $1.00 per million input tokens and $5.00 per million output tokens, according to Anthropic's own price page. We track its price at 4 providers. At a mix of three input tokens to one output token, it costs more than 71% of the 331 priced models we track.

How many benchmarks has Claude Haiku 4.5 been tested on?

We track 81 results for Claude Haiku 4.5 on 57 benchmarks from 22 sources, 27 of them independently verified. The latest was recorded on Oct 8, 2026.

Which API features does Claude Haiku 4.5 support?

OpenRouter lists tool calling, structured outputs, and reasoning for Claude Haiku 4.5.

About this record

Where Claude Haiku 4.5's numbers come from, and every name it appears under.

Tracked since
Apr 27, 2026
Newest source mention
Sep 7, 2026

Where the results come from

Verification: 81 scores · 27 independently verified · 34 aggregator-attributed · 1 vendor cross-reference · 19 vendor-reported. How these tiers are assigned

From 22 sources on 12 sites. Artificial Analysis supplies 34 of them; the 27 independently verified results come from 7 sites. Bars are coloured by trust tier.

  • artificialanalysis.ai34
  • api.llm-stats.com9
  • arcprize.org9
  • storage.googleapis.com6
  • www-cdn.anthropic.com6
  • anthropic.com4
  • raw.githubusercontent.com4
  • labs.scale.com3
  • epoch.ai2
  • swebench.com2
  • datasets-server.huggingface.co1
  • mistral.ai1

Also known as

claude 4.5 haiku (high reasoning)claude haiku 4.5 (thinking 16k)claude-haiku-4.5 (non-reasoning)claude haiku 4.5 (thinking 1k)claude-4-5-haikuclaude 4.5 haikuclaude 4.5 haiku (non-reasoning)claude-haiku-4-5-20251001haiku-4-5Haiku-4.5claude haiku 4.5 (thinking 32k)claude-haiku-4.5 (Reasoning)claude 4.5 haiku (reasoning)claude-haiku-4-5-20251001-thinkingclaude haiku 4.5 (thinking 8k)anthropic claude haiku 4.5and 3 more