Where is the Claude Haiku 5.5 documentation?

The Claude Haiku 5.5 documentation on OpenRouter includes the Claude 5.5 Migration Guide cookbook, the Reranking in RAG: How to Fix Bad Retrieval with a Reranker blog post, the How to Benchmark LLMs blog post, the How to Stop an AI Agent Stuck in an Infinite Loop blog post, the MCP vs CLI vs Skills: What Each Costs in Context blog post, the LLM API Rate Limits: How to Fix OpenAI and Claude 429 Errors blog post, the How to Get Structured Output from Any LLM: JSON Schema, Tool Calls, or a Decision Model blog post, the Best OCR Model for Scanned Documents, Measured blog post, the Build an AI Code Reviewer with OpenRouter blog post, the How to Build Your Own Always-On AI Agent blog post, the Claude Opus vs Sonnet vs Haiku: Which Should You Use? blog post, the How Much Does Claude Code Cost? Real Session Costs by Model blog post, the How to Extract Tables from a PDF with an LLM blog post, the Hit Your Claude Code Usage Limit? How to Keep Coding blog post, and the Claude 529 Overloaded: What It Means and How to Fix It blog post. The quickstart covers API keys and your first request, and the API reference lists every parameter.

How much does Claude Haiku 5.5 cost?

Claude Haiku 5.5 costs $0.10/M input tokens and $0.50/M output tokens, with separate rates for Cache Read at $0.01/M tokens, Cache Write at $0.125/M tokens, and Cache Write (1h) at $0.20/M tokens.

What is the context length of Claude Haiku 5.5?

Claude Haiku 5.5 has a 1,000,000 token context window. It supports up to 128,000 completion tokens.

Does Claude Haiku 5.5 support tool calling and structured outputs?

Yes. Claude Haiku 5.5 accepts tools and tool_choice for function calling. It also supports structured outputs via a JSON schema in response_format.

What inputs and outputs does Claude Haiku 5.5 support?

Claude Haiku 5.5 accepts text, images, and files such as PDFs as input and returns text.

Which providers serve Claude Haiku 5.5?

Claude Haiku 5.5 is served by 5 providers on OpenRouter: Google Vertex, Claude Platform on AWS, Azure, Anthropic, and Amazon Bedrock. Requests are routed to the best available provider, with automatic failover to the others, and you can pin or exclude providers with provider routing.

When was Claude Haiku 5.5 released?

Claude Haiku 5.5 was released on October 7, 2026.

Anthropic: Claude Haiku 5.5

anthropic/claude-haiku-5.5

Compare

Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and knowledge work, and accepts text and image input with a 1M-token context window.

It is the first Haiku model with adjustable effort. Thinking is adaptive and on by default, with effort as the main lever for trading off depth, latency, and cost; it can be turned off at low, medium, and high effort.

Modalities

In / out price

$0.10 / $0.50per 1M

Context

1.0M

Released

Oct 7, 2026

Providers

Different companies host the same model. OpenRouter routes your request to one of them based on the routing mode you pick: Balanced (price + speed), Nitro (fastest), or Floor (cheapest).

Latency / throughput


	$0.10	$0.50	$0.01	0.72s	170 tok/s	100.00%
	$0.10	$0.50	$0.01	1.24s	164 tok/s	99.48%
	$0.10	$0.50	$0.01	0.58s	154 tok/s	100.00%
	$0.10	$0.50	$0.01	1.44s	161 tok/s	98.94%
	$0.10	$0.50	$0.01	2.57s	145 tok/s	100.00%
Not used in Standard routing:Why these endpoints are not used
	$0.11	$0.55	$0.011	0.48s	166 tok/s	100.00%
	$0.11	$0.55	$0.011	0.52s	144 tok/s	99.97%
	$0.11	$0.55	$0.011	0.67s	165 tok/s	100.00%
	$0.11	$0.55	$0.011	0.71s	72 tok/s	100.00%
	$0.11	$0.55	$0.011	1.08s	136 tok/s	100.00%

Input $0.10Output $0.50
170 tok/s100.00% uptime
Input $0.10Output $0.50
164 tok/s99.48% uptime
Input $0.10Output $0.50
154 tok/s100.00% uptime
Input $0.10Output $0.50
161 tok/s98.94% uptime
Input $0.10Output $0.50
145 tok/s100.00% uptime
Not used in Standard routing:Why these endpoints are not used
Input $0.11Output $0.55
166 tok/s100.00% uptime
Input $0.11Output $0.55
144 tok/s99.97% uptime
Input $0.11Output $0.55
165 tok/s100.00% uptime
Input $0.11Output $0.55
72 tok/s100.00% uptime
Input $0.11Output $0.55
136 tok/s100.00% uptime

Pricing

The average price customers actually pay for this model, next to the prices providers post. Caching and discounts mean the price actually paid is often well below the listed one.

Weighted average

Input

$0.08158/M tokens

Output

$0.5961/M tokens

Claude Haiku 5.5 — Price history

Price history

No effective price data for this selection

	$0.0540	$0.517	66.7%	7.8%
	$0.0744	$0.579	61.0%	6.7%
	$0.0782	$0.571	59.7%	41.8%
	$0.0858	$0.609	54.0%	16.7%
	$0.0942	$0.660	55.2%	23.3%
Not used in Standard routing:Why these endpoints are not used
	$0.0625	$0.549	57.5%	0.0%
	$0.0882	$0.905	69.9%	2.8%
	$0.0927	$0.725	73.8%	0.1%
	$0.0984	$0.599	42.5%	0.4%
	$0.1041	$1.168	68.6%	0.3%

Effective in $0.0540Effective out $0.517
Token share
7.8%
Effective in $0.0744Effective out $0.579
Token share
6.7%
Effective in $0.0782Effective out $0.571
Token share
41.8%
Effective in $0.0858Effective out $0.609
Token share
16.7%
Effective in $0.0942Effective out $0.660
Token share
23.3%
Not used in Standard routing:Why these endpoints are not used
Effective in $0.0625Effective out $0.549
Token share
0.0%
Effective in $0.0882Effective out $0.905
Token share
2.8%
Effective in $0.0927Effective out $0.725
Token share
0.1%
Effective in $0.0984Effective out $0.599
Token share
0.4%
Effective in $0.1041Effective out $1.168
Token share
0.3%

Performance

Throughput is how fast the model writes (tokens per second — higher is better). Latency is total round-trip time (lower is better). TTFT is time-to-first-token — how long before you see anything appear (lower is better).

Throughput

170tok/s

P50, best across providers

Latency

0.48s

P50, best provider

Reasoning effortLatency / throughput

AutoExacto benchmarks

GPQA DiamondTAU-BenchAmazon Bedrock (US)87.4%74.3%Google Vertex (US)86.8%73.9%Google Vertex87.3%72.9%Amazon Bedrock86.7%73.2%Azure87.1%72.5%Amazon Bedrock (EU)86.7%72.5%auto-routing86.9%71.8%Azure (US)86.4%71.9%Claude Platform on AWS86.8%71.3%Google Vertex (EU)86.1%71.5%

+1 more providers

Uptime

Uptime is the percentage of the past 3 days that at least one provider was responding to requests. Availability is the percentage of time that inference was successfully served. OpenRouter continuously monitors and uses the next-best provider when one returns an error.

Uptime (3d)

99.99%

Availability (3d)

99.84%

Availability over the last 3 days

Last 72 hours

Availability 99.84%

3 Days Ago2 Days AgoYesterdayNow

Availability over the last 24 hours

OpenRouter Availability

99.65%

Without Routing

97.43%

When an error occurs in an upstream provider, we can recover by routing to another healthy provider, if your request filters allow it. You can access per-provider uptime data programmatically through the Endpoints API. Learn more about our load balancing and provider routing options.

Benchmarks

Scores on standardized evaluations and this model’s own outputs on shared benchmark prompts. Higher percentages are better — and each rank shows where this model lands among the models evaluated on that source.

Benchmark score summary for Anthropic: Claude Haiku 5.5 (Artificial Analysis, Design Arena and Vals AI)
Source	Benchmark	Score
Artificial Analysis	Claude Haiku 5.5 (Low) Intelligence Index	18.6
Artificial Analysis	Claude Haiku 5.5 (Low) HLE	27.0%
Artificial Analysis	Claude Haiku 5.5 (Low) AA-LCR	72.0%
Artificial Analysis	Claude Haiku 5.5 (Low) GDPval-AA	31.5%
Artificial Analysis	Claude Haiku 5.5 (Low) CritPt	9.1%
Artificial Analysis	Claude Haiku 5.5 (Low) SciCode	49.2%
Artificial Analysis	Claude Haiku 5.5 (Low) Terminal-Bench 4.0	12.6%
Artificial Analysis	Claude Haiku 5.5 (Low) AA-Omniscience Accuracy	30.4%
Artificial Analysis	Claude Haiku 5.5 (Low) AA-Omniscience Non-Hallucination Rate	59.5%
Artificial Analysis	Claude Haiku 5.5 (High) Intelligence Index	29.9
Artificial Analysis	Claude Haiku 5.5 (High) HLE	37.3%
Artificial Analysis	Claude Haiku 5.5 (High) AA-LCR	77.3%
Artificial Analysis	Claude Haiku 5.5 (High) GDPval-AA	46.2%
Artificial Analysis	Claude Haiku 5.5 (High) CritPt	18.6%
Artificial Analysis	Claude Haiku 5.5 (High) SciCode	48.7%
Artificial Analysis	Claude Haiku 5.5 (High) Terminal-Bench 4.0	21.7%
Artificial Analysis	Claude Haiku 5.5 (High) AA-Omniscience Accuracy	31.8%
Artificial Analysis	Claude Haiku 5.5 (High) AA-Omniscience Non-Hallucination Rate	61.1%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) Intelligence Index	34.6
Artificial Analysis	Claude Haiku 5.5 (Xhigh) HLE	42.7%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) AA-LCR	78.3%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) GDPval-AA	51.1%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) CritPt	22.6%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) SciCode	51.7%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) Terminal-Bench 4.0	29.3%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) AA-Omniscience Accuracy	33.8%
Artificial Analysis	Claude Haiku 5.5 (Xhigh) AA-Omniscience Non-Hallucination Rate	60.8%
Artificial Analysis	Claude Haiku 5.5 (Max) Intelligence Index	36.5
Artificial Analysis	Claude Haiku 5.5 (Max) HLE	44.4%
Artificial Analysis	Claude Haiku 5.5 (Max) AA-LCR	82.7%
Artificial Analysis	Claude Haiku 5.5 (Max) GDPval-AA	56.4%
Artificial Analysis	Claude Haiku 5.5 (Max) CritPt	18.9%
Artificial Analysis	Claude Haiku 5.5 (Max) SciCode	55.0%
Artificial Analysis	Claude Haiku 5.5 (Max) Terminal-Bench 4.0	32.8%
Artificial Analysis	Claude Haiku 5.5 (Max) AA-Omniscience Accuracy	34.7%
Artificial Analysis	Claude Haiku 5.5 (Max) AA-Omniscience Non-Hallucination Rate	65.8%
Artificial Analysis	Claude Haiku 5.5 (Medium) Intelligence Index	24.6
Artificial Analysis	Claude Haiku 5.5 (Medium) HLE	33.8%
Artificial Analysis	Claude Haiku 5.5 (Medium) AA-LCR	77.3%
Artificial Analysis	Claude Haiku 5.5 (Medium) GDPval-AA	39.1%
Artificial Analysis	Claude Haiku 5.5 (Medium) CritPt	12.9%
Artificial Analysis	Claude Haiku 5.5 (Medium) SciCode	49.0%
Artificial Analysis	Claude Haiku 5.5 (Medium) Terminal-Bench 4.0	15.2%
Artificial Analysis	Claude Haiku 5.5 (Medium) AA-Omniscience Accuracy	31.7%
Artificial Analysis	Claude Haiku 5.5 (Medium) AA-Omniscience Non-Hallucination Rate	59.6%
Design Arena	Claude Haiku 5.5 Models Arena 3D Elo	1321
Design Arena	Claude Haiku 5.5 Models Arena Asciiart Elo	1215
Design Arena	Claude Haiku 5.5 Models Arena Code Categories Elo	1276
Design Arena	Claude Haiku 5.5 Models Arena Data Visualization Elo	1279
Design Arena	Claude Haiku 5.5 Models Arena Game Development Elo	1314
Design Arena	Claude Haiku 5.5 Models Arena SVG Elo	1194
Design Arena	Claude Haiku 5.5 Models Arena UI Component Elo	1275
Design Arena	Claude Haiku 5.5 Models Arena Website Elo	1252
Vals AI	Claude Haiku 5.5 BioMysteryBench	70.4%
Vals AI	Claude Haiku 5.5 ClaimBench	100.0%
Vals AI	Claude Haiku 5.5 Code Migration v1.1	52.4%
Vals AI	Claude Haiku 5.5 CyberBench v1.1	75.5%
Vals AI	Claude Haiku 5.5 EMB v1.1	65.5%
Vals AI	Claude Haiku 5.5 Finance Agent v2.1	65.4%
Vals AI	Claude Haiku 5.5 Vals Finance Index	57.4%
Vals AI	Claude Haiku 5.5 HLAB-Vals-1.1	4.2%
Vals AI	Claude Haiku 5.5 Immigration	61.8%
Vals AI	Claude Haiku 5.5 IOI	47.3%
Vals AI	Claude Haiku 5.5 Vals Legal Index	70.3%
Vals AI	Claude Haiku 5.5 Legal Research Bench	43.8%
Vals AI	Claude Haiku 5.5 MedCode	51.5%
Vals AI	Claude Haiku 5.5 MedScribe	89.5%
Vals AI	Claude Haiku 5.5 MysteryMechanism	35.1%
Vals AI	Claude Haiku 5.5 ProgramBench	2.5%
Vals AI	Claude Haiku 5.5 ProofBench v1.1	89.0%
Vals AI	Claude Haiku 5.5 Public Benefits Bench v1.1	63.2%
Vals AI	Claude Haiku 5.5 SAGE	89.8%
Vals AI	Claude Haiku 5.5 Vals Science Index	38.5%
Vals AI	Claude Haiku 5.5 SRE Bench	5.7%
Vals AI	Claude Haiku 5.5 SWE-bench Multimodal	44.0%
Vals AI	Claude Haiku 5.5 Tax Agent Bench	31.9%
Vals AI	Claude Haiku 5.5 Terminal-Bench 4.0	36.4%
Vals AI	Claude Haiku 5.5 Terminal-Bench-Science	10.0%
Vals AI	Claude Haiku 5.5 Vals Index	58.7%
Vals AI	Claude Haiku 5.5 Vibe Code Bench 1-100	20.4%
Vals AI	Claude Haiku 5.5 Vibe Code Bench v1.1	90.4%

Apps

Public apps that send the most traffic to this model. Good signal for what real production workloads look like — and a hint at which use cases this model is best suited for.

Claude Code

Claude Code is Anthropic's agentic coding tool that reads your entire codebase, plans and executes changes across files, runs tests, and iterates on failures, all from natural language prompts.

780Btokens

Hermes Agent

Hermes Agent is an open-source, self-improving AI agent by Nous Research that runs persistently with memory across sessions, and builds reusable skills from experience. It comes with 40+ built-in tools, including web search, browser automation, and vision, plus scheduled automations and subagents.

461Btokens

There are many coding agents, but this one is yours.

157Btokens

Freebuff

Powerful coding models, funded by ads

138Btokens

Codex

A coding agent that helps you build and ship with AI

105Btokens

Rest of OpenRouter

297B

Top 5 Apps

51.3B

Anthropic: Claude Haiku versions

Available through EU in-region routing

Anthropic: Claude Haiku 5.5 is available through EU in-region routing. Requests sent to eu.openrouter.ai are decrypted in the EU and routed only to shared provider endpoints in Europe (the EU/EEA, the UK, and Switzerland).

Providers in Europe

Google Vertex, Amazon Bedrock

Prices in Europe

Google Vertex: $0.11/M tokens input, $0.55/M tokens output
Amazon Bedrock: $0.11/M tokens input, $0.55/M tokens output

Base URL

https://eu.openrouter.ai/api/v1

Plan

Requires Business or Enterprise. Self-serve on Business: no contract, no monthly minimum. OpenRouter Business for EU in-region routing

Fails closed: if no shared provider endpoint in Europe is available, the request errors instead of leaving the region. Requests served by your own BYOK key go to your deployment’s region, which you control. In-region routing docs

More models from Anthropic

Claude Fable 5.2

Claude Fable 5.2 is Anthropic's most capable model for the most demanding technical work, with gains over Fable 5.1 in coding, reasoning, and vision. In prerelease testing it completed otherwise-infeasible software migrations, designed working electronic devices from scratch, and made progress on open problems in physics and mathematics. It communicates more clearly than earlier Fable models and generates output up to 30% faster than Fable 5.1. Adaptive extended thinking is always on, with reasoning effort levels from low to max (default high) and preserved thinking across turns. Anthropic recommends Opus 5.5 as the daily driver for most coding and knowledge work, with Fable 5.2 for the most ambitious, open-ended problems.

Text1M context$8 / $40

Claude Fable 5.2 (batch)

Text1M context$4 / $20

Claude Haiku 5.5 (batch)

Text1M context$0.05 / $0.25

Claude Sonnet 5.5

Claude Sonnet 5.5 is Anthropic's Sonnet-class model for well-scoped everyday work, succeeding Claude Sonnet 5 as a direct upgrade. It is especially strong at building features, fixing bugs, and producing polished documents, slides, and spreadsheets, and it writes and communicates more clearly than its predecessor.

Thinking is always on, so effort is the main lever for trading off depth, latency, and cost, and lower effort settings keep it responsive for everyday agentic loops.

Text1M context$2 / $10

Claude Sonnet 5.5 (batch)

Thinking is always on, so effort is the main lever for trading off depth, latency, and cost, and lower effort settings keep it responsive for everyday agentic loops.

Text1M context$1 / $5

Claude Opus 5.5

Claude Opus 5.5 is Anthropic's flagship model for demanding reasoning, coding, and long-horizon agentic work, succeeding Claude Opus 5. It is particularly strong at multi-step changes in large codebases, code review and bug finding, financial and scientific analysis, and reading dense charts, diagrams, and screenshots, and it is more careful than its predecessor about only stating figures and citing sources it can back up.

The model completes comparable tasks in fewer steps and with fewer tokens than Opus 5, and reports on its work in plainer language, with clear updates on what it did, what it found, and what it needs from the user. Thinking is always adaptive, so effort is the main lever for trading off depth, latency, and cost, and lower effort settings remain effective for latency-sensitive workloads.

Text1M context$4 / $20

Claude Opus 5.5 (batch)

Text1M context$2 / $10

Claude Fable 5.1

Claude Fable 5.1 improves on Claude Fable 5 across the board, with the biggest gains in agentic coding, long-running agentic workflows, and knowledge work: long code refactors, front-end and visual code generation, and finance and analysis tasks in particular. It also tends to be more concise than Fable 5 in its plans and summaries. We recommend testing it as a direct upgrade wherever you use Fable 5 today, and alongside Opus 5 on reasoning-heavy tasks.

Text1M context$10 / $50

Claude Fable 5.1 (batch)

Text1M context$5 / $25

Claude Opus 5

Fast-mode variant of Opus 5 - identical capabilities with higher output speed at 2x pricing relative to regular Opus 5.

Learn more in Anthropic's docs: https://platform.claude.com/docs/en/build-with-claude/fast-mode

Note: As of September 1, 2026, this dedicated fast model is deprecated. Fast mode is now served by the fast service tier endpoint on Claude Opus 5 — request it with service_tier: "fast" or speed: "fast". Requests to this model keep working and are served by the same fast tier capacity, so no action is required, but new integrations should target the regular model. Learn more in our service tier docs: https://openrouter.ai/docs/guides/features/service-tiers

Text1M context

Claude Opus 5

Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work. It is particularly strong at end-to-end software tasks, code review and bug finding, visual analysis of charts and documents, complex office deliverables, and coordinating parallel subagents.

The model maintains strong instruction following and tool use across extended tasks, while remaining effective at lower effort settings for workloads that prioritize latency and token efficiency.

Text1M context$5 / $25

Claude Opus 5 (batch)

The model maintains strong instruction following and tool use across extended tasks, while remaining effective at lower effort settings for workloads that prioritize latency and token efficiency.

Text1M context$2.50 / $12.50

Claude Sonnet 5

Sonnet 5 is Anthropic's most capable Sonnet-class model, with frontier performance across coding, agents, and professional work. It supports adaptive thinking with selectable reasoning effort levels (low, medium, high, max, and x-high), a 1M-token context window, and text, image, and file inputs. Sonnet 5 uses an updated tokenizer and includes real-time cyber safeguards that block certain high-risk dual-use activities.

Text1M context$2 / $10

Claude Sonnet 5 (batch)

Text1M context$1 / $5

Claude Fable Latest

This model always redirects to the latest model in the Claude Fable family.

Text1M context

Claude Fable 5

Claude Fable 5 is a Mythos-class model from Anthropic, built for autonomous knowledge work and coding. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token context window. It is suited for long-running, complex, and asynchronous tasks that previously required frequent human check-ins.

It is particularly strong at end-to-end work that would otherwise take a person hours, days, or weeks - taking on problems that are long-running, ambiguous, or highly multi-step. It executes well-scoped tasks with few mistakes, automatically self-correcting through verification loops, and ships with robust safeguards.

Text1M context$10 / $50

Claude Fable 5 (batch)

Text1M context$5 / $25

Claude Opus 4.8

Fast-mode variant of Opus 4.8 - identical capabilities with higher output speed at 2x pricing relative to regular Opus 4.8.

Learn more in Anthropic's docs: https://platform.claude.com/docs/en/build-with-claude/fast-mode

Note: As of September 1, 2026, this dedicated fast model is deprecated. Fast mode is now served by the fast service tier endpoint on Claude Opus 4.8 — request it with service_tier: "fast" or speed: "fast". Requests to this model keep working and are served by the same fast tier capacity, so no action is required, but new integrations should target the regular model. Learn more in our service tier docs: https://openrouter.ai/docs/guides/features/service-tiers

Text1M context

Claude Opus 4.8

Claude Opus 4.8 is Anthropic's most capable generally available model in the Opus family. It supports text, image, and file inputs with text output, with reasoning support and a 1M-token context window. It is suited for highly autonomous agents, long-horizon agentic work, knowledge work, and memory-driven tasks where coherence over extended sessions matters.

It is particularly strong on multi-step reasoning, complex coding, and end-to-end project orchestration - large codebases, multi-stage debugging, and long-running asynchronous agent pipelines. Beyond coding, it handles knowledge work such as drafting documents, building presentations, and analyzing data, maintaining quality across very long outputs.

Text1M context$5 / $25

Claude Opus 4.8 (batch)

Text1M context$2.50 / $12.50

Anthropic: Claude Haiku 5.5

Providers