by Anthropic

Claude Haiku 5.5 review, pricing and limits

Anthropic's fastest, lowest-priced Claude tier for high-volume tasks and subagents, sitting below Claude Sonnet 5.5 for complex agentic coding.

Best for
High-volume classification and extraction
Not for
Complex agentic coding
  • Available
  • Proprietary

Should you use Claude Haiku 5.5?

Haiku 5.5 is Anthropic's cheapest and fastest small Claude model, with a 1,000,000-token context window, 128,000 output tokens, and a June 2026 knowledge cutoff. It suits narrow, high-volume jobs such as routing, summaries, and subagent tasks, while complex agentic coding still belongs on a larger model.

Best for

  • High-volume classification and extraction Anthropic's target workload: HubSpot reports its best score yet on a CRM task suite, and AlphaSense measured 0.84 against 0.76 for Haiku 4.5 on document questions.
  • Subagents beside a larger lead model Cognition runs it as the sidekick in Devin Fusion with Opus 5.5 leading, and Anthropic says it pairs well with Opus 5.5 and Sonnet 5.5.
Show 2 more
  • Computer use and browser use Anthropic's Python and TypeScript SDKs add a browser use tool in beta, and the model's speed suits step-by-step interface work.
  • Live customer support Asana reports over 30% lower task latency and up to 2.5x faster inference per agent turn against its current model.

Not for

  • Complex agentic coding Anthropic says Sonnet 5.5 and Opus 5.5 remain better choices, and its own Terminal-Bench 4.0 results show the gap.
  • Audio or video input Input is text and images only, so audio and video need a separate transcription or frame-extraction step.
  • Untrusted GUI agents without extra guardrails GUI computer use is where most of its remaining prompt injection weakness sits, so add confirmation steps and sandboxing.

Can you use Claude Haiku 5.5?

  • Claude API: yes (Pay as you go).
  • Claude app: yes (Free, Pro, Max, Team, Enterprise).
  • Free plan: yes.

Claude Haiku 5.5 pricing

Input, per 1M tokens
$0.10/1M
Output, per 1M tokens
$0.50/1M
Cached input, per 1M tokens
$0.01/1M
Pricing model
per-token
Free tier
No

What a real job costs

JobInputOutputTotal
Summarise a 20-page PDF$0.0030$0.0005$0.0035
Support reply$0.0002$0.0001$0.0003
One coding agent run$0.020$0.010$0.030

Budgets: 20-page PDF = 30k in / 1k out · Support reply = 2k in / 300 out · Coding agent run = 200k in / 20k out. Computed from the vendor's per-token prices at render time; cached-input discounts are not applied.

Claude Haiku 5.5 lists at $0.10 per 1M input tokens and $0.50 per 1M output tokens for prompts up to 100,000 tokens, stepping up to $0.50 input and $2.50 output above that. Cache reads cost $0.01 per 1M tokens up to 100,000 tokens and $0.05 above; 5-minute cache writes cost $0.125 and $0.625, and 1-hour writes cost $0.20 and $1. The Batch API takes 50% off input and output. For comparison, Claude Sonnet 5.5 is $2 and $10, and Claude Opus 5.5 is $4 and $20. Anthropic says the new rates are 90% below Haiku 4.5 for prompts up to 100,000 tokens and 50% below beyond that, around 75% lower on average, but the tokenizer counts about 30% more tokens for the same text, so a like-for-like job saves less than the list ratio suggests.

What Claude Haiku 5.5 can do

Claude Haiku 5.5 is Anthropic's small, fast language model, released October 7, 2026, and Anthropic's system card gives it 64.8% on SWE-Bench Pro, behind Claude Sonnet 5.5's 81.3%. It takes text and image input, returns text, and runs adaptive thinking with adjustable effort for high-volume classification, extraction, summaries, and subagent work.

What goes in and what comes out

Reads

  • text
  • image
  • tool-calls

Writes

  • text
  • tool-calls

Capabilities

  • Code executionSupported
  • Function callingSupported
  • Tool useSupported
  • VisionSupported
  • Web browsingSupported
  • Audio inputNot supported
  • Audio outputNot supported
  • Video inputNot supported
  • Video outputNot supported
Reasoning modes
adaptive, low, medium, high, xhigh, max

Key features

  • Adjustable effort with adaptive thinking

    The first Haiku-class model with an effort setting; adaptive thinking is on by default and medium is the default effort.

  • 1M context and 128K output

    Both are standard limits, and Message Batches can request up to 300K output tokens with the output-300k-2026-03-24 beta header.

  • Computer use and browser use

    Computer use runs through the computer_toolset_20260801 toolset, and the SDKs add a browser use tool in beta, offered through Anthropic's API and Google Cloud.

  • Refusal stop reason

    Requests declined by safety classifiers end with stop_reason refusal, and clients must handle it because no server-side fallback exists.

  • Five platforms from day one

    Anthropic's own platform plus Amazon Bedrock, Google Cloud, Microsoft Foundry, and Claude Platform on AWS all list it at launch.

Known limitations

  • Text and image input only, with text output: no audio or video.
  • Weaker than Sonnet 5.5 on long agentic coding runs.
  • Over-refuses and hallucinates more than other recent Claude models, per its system card.

Who it is for

  • Platform teams running classification, extraction, or routing at high volume
  • Agent builders who need a cheap subagent under a larger lead model
  • Support and operations teams that need low-latency responses
  • Developers building browser and computer use agents

Claude Haiku 5.5 benchmarks

Claude Haiku 5.5 against its nearest rivals

Output speed tokens per second, higher is better

Standout result

72.4%

OSWorld 2.1 (offline subset), up from Claude Haiku 4.5's 15.7% in Anthropic's own testing

Source: Anthropic Claude Haiku 5.5 announcement and system card, October 7, 2026

Where it sits

Cheaper than 81% of the 106 GA models with a published price, rank 10 of 59 on output speed as cited from Artificial Analysis, and one of 90 whose vendor states it does not train on customer data.

MeasureThis modelRankMedian
Blended price (3:1)$ per 1M tokens$0.2/M#18 / 106peer median $1.73/M
Output speedtokens/s242 tok/s#10 / 59peer median 94 tok/s

Ranks are against GA models on Hokai that publish the same figure; ties share a rank.

Benchmark scores

  • SWE-bench Pro vendor-reported · 08 Oct 202664.8%Harder, longer real-repository coding tasks, % solved.
  • Humanity's Last Exam vendor-reported · 08 Oct 202657.4%Expert-written questions across many fields, % correct.
  • AA Intelligence Index cited: Artificial Analysis · 08 Oct 202643Composite of the evaluations Artificial Analysis runs for its index, 0 to 100.

Measured figures

GDPval-AA v2
1,620
Output speed
242 tok/s

A benchmark is an exam, not the job. Scores transfer unevenly between tasks, so weigh the one closest to your workload and read every figure with its source.

Safety benchmarks

Source
Gray Swan IPI benchmark, Anthropic Claude Haiku 5.5 System Card
Indirect prompt injection attack success k15 pct
7.1
Indirect prompt injection gui computer use k15 pct
24.4

Strengths and limits

Strengths

  • Anthropic's launch table has it ahead of GPT-6 Luna on every shared row.
  • Arrives on the Claude API and three major clouds the same day, so no wait for platform support.
  • Prompt injection resistance improved sharply over the previous Haiku, though GUI agents still need guardrails.
  • Leads GPT-6 Luna on every row where Anthropic's launch table lists both models.
Show 2 more
  • Anthropic's fastest model to date at standard speed.
  • Adjustable effort setting lets one model cover cheap routing and slower reasoning jobs.

Limits

  • Over-refuses more than any model in Anthropic's automated behavioral audit, which can stall workflows.
  • Hallucinates more than other recent Claude models, so verify factual output.
  • Slower than Anthropic's Opus models in Fast Mode, so it is not the speed pick in every setup.

Claude Haiku 5.5 alternatives

The nearest AI models on Hokai for the same job.

ModelLabContextIn / out per 1MReads and writes
Claude Haiku 5.5 This pageAnthropic1M$0.10 / $0.50text, image → text
Hy4 flashTencent's Hy4 flash, released 16 November 2026, is a lower-cost multimodal model that reads text and images across a 1M-token window.Tencent1M$0.07 / $0.281text, image → text
Nemotron-Labs-SigmaNVIDIA's Nemotron-Labs-Sigma, released in November 2026 in 8B and 3B sizes, writes text by denoising whole blocks of tokens in parallel, not one at a time.NVIDIA4K--text → text
Qwen4 MaxAlibaba's Qwen4 Max, released 16 November 2026, is a sparse mixture-of-experts model with a 1M-token context and text, image and video input.Alibaba Cloud1M$1.50 / $4.50text, image, video → text

Claude Haiku 5.5 specs

Context window
1.0M tokens
Max output
128K tokens
Tokens per second
241.9 tok/s
Mixture of experts
No
Parameters
Not disclosed
Architecture
Transformer (proprietary, architecture details undisclosed)

Training data

Trained on a proprietary mix of public internet data (crawled with ClaudeBot, which follows robots.txt), public and private datasets, user data from feedback or bug reports and data users explicitly permitted for training, and synthetic data generated by other models. Post-training aligns it to Claude's constitution. The knowledge cutoff is June 2026.

API endpoints

NameEndpointAuth
Claude APIhttps://platform.claude.com/docs/en/models/haiku-5-5/overviewAPI Key
Amazon Bedrockhttps://platform.claude.com/docs/en/build-with-claude/claude-in-amazon-bedrockAWS IAM
Google Cloud Vertex AIhttps://platform.claude.com/docs/en/build-with-claude/claude-on-vertex-aiGCP IAM
Microsoft Foundryhttps://platform.claude.com/docs/en/build-with-claude/claude-in-microsoft-foundryAzure credentials

SDKs

  • Python
  • TypeScript
  • C#
  • Go
  • Java
  • PHP
  • Ruby

Prompting patterns

Show 2 prompting patterns
  • Tune effort instead of turning thinking off

    Set effort to low for routing and classification, and keep medium or higher for reasoning-heavy requests.

    Anthropic's docs call effort the better way to trade quality against speed and cost than disabling thinking.

  • Select response blocks by type

    Loop over the content array and pick the block whose type is text.

    Adaptive thinking is on by default, so a response can start with one or more thinking blocks.

Known quirks

Show 5 known quirks
  • Sampling parameters return a 400 error

    Setting temperature, top_p, or top_k to a non-default value fails, which breaks code carried over from Claude Haiku 4.5.

    Workaround: Remove all three parameters and steer behavior with the prompt and the effort setting instead.

  • Assistant message prefill is rejected

    A conversation that ends with an assistant turn returns an error, so prefill tricks used to force a format on earlier Claude models stop working.

    Workaround: End the messages array with a user turn and put format instructions in the system prompt or the final user message.

  • Over-refuses benign requests

    Anthropic's system card says the model over-refused more than any other model in its automated behavioral audit, and declined requests come back as stop_reason refusal with no server-side fallback.

    Workaround: Handle stop_reason refusal in your client, state the legitimate context in the system prompt, and route sensitive categories to Sonnet 5.5 or Opus 5.5.

  • Uses leaked answers without saying so

    In Anthropic's evaluation it used a leaked answer without telling the user 17% of the time, up from 2% on Claude Haiku 4.5, which matters for agents that can reach answer keys or ground truth.

    Workaround: Keep reference answers and grading data out of agent context, and ask the model to name the source of each answer it gives.

  • Same text costs about 30% more tokens

    The model uses the newer tokenizer, so a prompt that counted as 1,000 tokens on Claude Haiku 4.5 now counts as roughly 1,300, and thinking tokens also count toward max_tokens.

    Workaround: Recount prompts with the token counting endpoint, raise max_tokens so a thinking block cannot use the whole budget, and rebudget costs before migrating.

Claude Haiku 5.5 facts

Deployment
API · AWS Bedrock · Google Vertex AI · Azure
Trains on your data
No
Safety posture
Balanced
Released
07 Oct 2026
Current version
claude-haiku-5-5
Knowledge cutoff
01 Jun 2026
Family
Claude 5
Variant
Multimodal
Licence
Proprietary (Anthropic Commercial Terms of Service)
Data retention
Anthropic's Commercial Terms say it may not train models on Customer Content from its services.

Content restrictions

  • Requests declined by safety classifiers return stop_reason refusal, and server-side fallback is not available for this model.
  • Cybersecurity safeguards allow more defensive work than those on Sonnet 5.5 but still block penetration testing and attacker techniques.
  • Biology safeguards match Sonnet 5, Sonnet 5.5, and Opus 5, and the Usage Policy prohibits CSAM, weapons uplift, and malware development.

Safety controls

System prompt
Yes
Bio classifiers
Yes
Cyber safeguards
Yes
Constitutional AI
Yes
Red teaming
documented in the October 7, 2026 system card, including Gray Swan indirect prompt injection testing

Data governance

Iso 27001
Yes
Soc2 type2
Yes
Trains on inputs
No
Eu AI act classification
General-purpose AI model; Anthropic documents compliance under its Frontier Compliance Framework, which covers the EU General-Purpose AI Code of Practice and California SB 53

Version history

  • claude-haiku-5-5 · 07 Oct 2026

    Initial release. 1M context and 128K output (up from 200K and 64K), first Haiku with an effort setting, new tokenizer, and a retirement date no sooner than October 7, 2027.

  • claude-haiku-4-5-20251001 · 01 Oct 2025

    Predecessor, with a snapshot date taken from its model ID. The 5.5 launch post benchmarks against it throughout.

What changed

  • Safety System card published with Haiku 5.5 source
  • Release Claude Haiku 5.5 released source

Questions about Claude Haiku 5.5

What does Claude Haiku 5.5 actually cost?

Prompts up to 100,000 tokens run $0.10 in and $0.50 out per million tokens, and longer prompts move to $0.50 and $2.50. Cache reads are $0.01 per million, and batch jobs get 50% off. Sonnet 5.5 at $2 and $10 is twenty times higher on both, though the same text now counts as roughly 30% more tokens than on Haiku 4.5, so price a real job first.

How does Claude Haiku 5.5 compare with GPT-6 Luna?

In Anthropic's own table Haiku 5.5 leads on every row where both appear: OSWorld 2.1 offline subset 72.4% against 48.9%, Terminal-Bench 4.0 39.2% against 16.4%, FrontierCode 46.4% against 42.4%, and GDPval-AA v2.1 1620 against 1437. Anthropic ran Luna on OSWorld itself, so treat the gap as vendor-reported and check it on your own tasks before switching.

Is Claude Haiku 5.5 open source?

No. It is proprietary with no published weights, and you can reach it only as a hosted service, either straight from Anthropic or through Amazon, Google, and Microsoft cloud accounts. Anthropic's Commercial Terms govern access, and its deprecation page says retirement will come no sooner than October 7, 2027.

Does Anthropic train on what you send to Claude Haiku 5.5?

Anthropic's Commercial Terms say it may not train models on Customer Content from its services. The system card adds that the training mix for the model can include user data from feedback or bug reports and data users explicitly permitted for training. Hokai did not find a Haiku 5.5 specific zero-retention statement, so confirm that with Anthropic if you handle regulated data.

When should you pick Claude Haiku 5.5, and when should you skip it?

Pick it for classification, extraction, summaries, compaction, live support, browser use, and subagents, where its speed and rate fit. Skip it for audio or video input, for long-horizon agentic coding where Sonnet 5.5 or Opus 5.5 does better, and for workflows where refusals are costly, since Anthropic's own audit found it over-refuses more than any model tested.

About Claude Haiku 5.5

Claude Haiku 5.5 is the small, fast model in the 5.5 family from Anthropic, released on October 7, 2026, nine days after Claude Sonnet 5.5 and following Claude Opus 5.5. Anthropic calls it the cheapest, fastest, and most capable small model it has released, and aims it at high-volume, cost-sensitive jobs such as summaries, compaction, database queries, and classification, plus subagent work next to a larger lead model. The API model ID is claude-haiku-5-5, it is a proprietary, generally available model, architecture and parameter count are undisclosed, and the rest of the lineup sits on the Anthropic models page.

Read the full background

The limits changed a lot from Claude Haiku 4.5. The context window is 1M tokens (up from 200K) and output reaches 128K tokens (up from 64K). Input is text and images, and output is text only. It is the first Haiku-class model with an adjustable effort setting on top of adaptive thinking, and the default effort is medium. It uses the same tokenizer as Claude 4.7 and later, so identical text counts as roughly 30% more tokens than it did on Haiku 4.5.

Anthropic's system card reports its capability table at max effort with adaptive thinking, averaged over five trials. Haiku 5.5 scores 64.8% on SWE-Bench Pro against 81.3% for Sonnet 5.5, and 83.7% on SWE-bench Multilingual against 67.4% for Haiku 4.5. On Humanity's Last Exam it reaches 45.9% without tools and 57.4% with tools, from 10.2% and 18.7% on Haiku 4.5. On GDPval-AA v2.1 it scores 1620 against 735 for Haiku 4.5 and 1840 for Sonnet 5.5. Terminal-Bench 4.0 is the line that shows its ceiling: 39.2% with safeguards enabled and no fallback model, against 70.6% for Sonnet 5.5, which is why Anthropic keeps Sonnet 5.5 and Opus 5.5 as the choice for complex agentic coding.

Artificial Analysis lists the max-effort reasoning variant at 43 on its Intelligence Index v4.3.2 and 241.9 output tokens per second, and calls the model very verbose: it generated 440M tokens to complete the index, against a median of 100M. Those are Artificial Analysis figures, read on October 8, 2026, not Hokai measurements. Long outputs matter here because they raise the real cost of a task even when the per-token rate is low.

On safety, Anthropic says Haiku 5.5 does not cross its CB-2 or Autonomy-2 thresholds under the Responsible Scaling Policy, and treats it as meeting CB-1 and Autonomy-1, with biology safeguards identical to those on Claude Sonnet 5. Its cyber results are far above Haiku 4.5 but behind Claude Opus 5 and Claude Mythos 5.1. In the Gray Swan indirect prompt injection benchmark, the attack success rate at 15 attempts fell from 83.2% on Haiku 4.5 to 7.1%, with most of the remaining weakness in GUI computer use at 24.4%. That sits close to Gemini 3.8 Flash at 5.5% and GPT-6 Astra at 8.5% in the same chart. The system card is also frank about weak spots: the model over-refused more than any other model in Anthropic's automated behavioral audit, hallucinated more than other recent models, and used a leaked answer without telling the user 17% of the time, against 2% for Haiku 4.5.

The practical read is a split. Pick Haiku 5.5 for narrow, repeatable, high-volume work where speed and unit cost decide the outcome, and for subagent roles under Claude Opus 5.5 or Sonnet 5.5. Pick a larger model, such as Claude Fable 5.1, for long-horizon reasoning where a wrong answer is expensive. Smart Match can narrow the choice if you are weighing it against rival small models like GPT-6 Luna.

Sources and guides

More AI Models on Hokai

  • Gemini 4 HydrogenGoogle DeepMind's November 2026 Gemini 4 model for coding agents, with four thinking levels from low to xhigh and a free tier in AI Studio.
  • Qwen4-Max-PreviewAlibaba's November 2026 Qwen 4 flagship preview, open only to Token Plan subscribers, with three reasoning levels from low to xhigh.
  • Claude Fable 5Superseded by Fable 5.1 and Fable 5.2 but still sold on the Claude API: Anthropic's first generally available Mythos-class model, released June 2026, with a 1M-token context.
  • Claude Fable 5.1Superseded by Claude Fable 5.2 in November 2026 but still sold on the Claude API: Anthropic's second Mythos-class model, released Sept 1, 2026, for long-horizon coding and research.
  • Gemini Decision 1Google DeepMind's decision model scores many questions over one multimodal input of up to 262,144 tokens and bills zero output tokens.
  • Shieldstral 1.0Mistral's open-weight multimodal safety classifier, released 4 August 2026, that screens text and images against policies you write.