Claude Haiku 5.5
Very cheap small model for summaries, sorting and coding subagents; use Sonnet 5.5 for hard coding work.
This product is made by Anthropic, whose Claude models help draft our reviews. How we keep that from tilting a score
Should you use it?
Pick it if
Developers running high-volume jobs (summaries, classification, extraction) or coding subagents on a tight budget.
Skip it if
You need near-top coding or agent work; Sonnet 5.5 ranks 3rd on WebDev, and the best non-Anthropic pick, GPT-6.1 Sol, 4th s1.
The short version
- What this is
- Anthropic's small, cheap model, released 7 Oct 2026, on the Claude API and on AWS, Google Cloud and Microsoft Azure, with an adjustable effort setting.
- Why it matters
- It codes about as well as some mid-size models for a fraction of their price, so bulk and background jobs get much cheaper.
- Rank 30 of 142 on arena.ai WebDev, which scores 44 for coding s1.
- API $0.10 / $0.50 per 1M tokens up to 100K-token prompts s3.
- Covered by Reuters and YourStory at launch s7 s8.
Good at
Our take
Haiku 5.5 is Anthropic's budget model, and the one public board that lists it puts it at rank 30 of 142 on WebDev s1. That scores 44 for coding, well short of the top. Its vote count is still under 1,000, so the rank may move.
The case for it is price. At $0.10 / $0.50 per 1M tokens s3 it matches GPT-6 Luna's token price and ranks a little above it on WebDev s1. Anthropic itself says to pick Sonnet 5.5 or Opus 5.5 for complex agentic coding s3. We'll add writing, research and agents scores once the arena.ai boards list it.
Score
A score from 0 to 100 per job, against everything available today. Overall 44 across 1 job, the average of the job scores.
New to this? What the scores and labels mean
- 0 to 100
- Our score for one job, like writing or coding, compared with everything else out today. 100 is the best available today. For AI models we build it from public data, such as Arena's blind votes and Epoch AI's independent tests; apps and tools are scored on our own checklist. The same tool can score differently for different jobs, and where two likely ranges overlap, we call it a tie.
- Our verdict
- Use it: a top choice for most people. Worth trying: good, and worth a go. Depends: good for some people or one kind of job. Wait: too early to tell, or not out for everyone yet. Skip: something else does the job better.
- First look
- Just released. We say what it is and whether to bother, and give it a score once public results exist.
- No score yet
- None of the public sources we use covers it yet, or the only evidence is from a source whose numbers we can't use, so we describe it in words instead.
- Best buy
- Nothing cheaper scores as well for that job.
- Best free
- The best score for that job that costs nothing, through a free plan or by running on your own computer.
- Pay as you go
- No monthly plan: you pay for each use, usually through a developer account.
- Checked
- The day we last confirmed the score and the price against the sources. Prices change often, so check the date.
- AI model
- The engine underneath, like GPT or Claude. You use it through an app or, if you code, directly.
- Open source
- Free to download and run on your own computer, so nothing leaves your machine.
#19 of the 21 we have scored for this job.
From blind votes (Arena) · Arena data, CC BY 4.0
1 source so farData read 20 Oct
Alternatives
Better or cheaper for coding, from our scores, then the ones we'd also look at. The same kind of product comes first.
- Claude Opus 5.5BetterText model100$20 a monthScores 100 in coding, against 44 here; $20 a month. Side by side
- Claude Sonnet 5.5FreeText model89FreeScores 89 in coding, with a free plan. Much stronger at coding and agents for twenty times the token price. Side by side
- GLM-5.3-FlashText model48FreeScores 48 for coding, against 44 here, with MIT weights you can host yourself; $0.15 / $0.50 on Z.ai. Side by side
Evidence
Everything behind the call and the scores, folded so the page stays short. Spotted a wrong price, score or link? Report a mistake.
How we scored this
The coding score is our own, built from public data: blind votes on Arena and, where they cover the job, Epoch AI's independent tests. How we combine them is our method and we don't publish it.
Public leaderboard
This product is made by Anthropic, whose Claude models help draft our reviews.
Reviews are drafted with help from AI models, including Anthropic's Claude. Model scores come only from outside leaderboards and independent evaluations, by a written rule.
No one paid for this review or gave us free access, and we earn nothing if you sign up. How we score
Sources 8
- s1arena.ai Code Arena: WebDev leaderboard
- s2arena.ai Text: Coding leaderboard
- s3Anthropic: Introducing Claude Haiku 5.5
- s4Artificial Analysis LLM leaderboard
- s5Claude privacy: is my data used for model training (commercial)
- s6arena.ai Agent leaderboard
- s7Deccan Chronicle (Reuters): Anthropic launches third Claude 5.5 model
- s8YourStory: Claude Haiku 5.5 debuts as Anthropic's faster small AI model
Something here out of date or wrong? Report a mistake and we will check it against the source.
History 2 changes
Every change to a rating, the price or our call, with the reason. Scores are against what exists on the day, so they drift down as better things ship. Older entries come from before we moved to scores.
- New: Coding ratedReviewed: coding. Page under the news test (Reuters and YourStory coverage); that rating alone does not meet the value test.
- Audit: added GLM-5.3-Flash as an alternative (same coding rating, open weights); dropped Gemini 3.8 Flash, which ranks lower and costs more. Rating unchanged.
Use it well 4 guides on this site
Pages on this site that help with the jobs above.