OpenRouter 模型 ↗ · 2026-10-08
anthropic/claude-haiku-5.5
模型发布 OpenRouter 模型 原文语言:中文
OpenRouter 于 2026-10-08 收录该模型(收录时间不等于模型正式发布日期);上下文:1,000,000 tokens;输入价:$0.1000/百万 tokens;输出价:$0.5000/百万 tokens;定位(来源描述):Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as…
全文
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
Compare
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$0.0101.09s166 tok/s
99.50%
$0.10$0.50$0.0101.43s170 tok/s
100.00%
$0.10$0.50$0.0101.34s167 tok/s
100.00%
$0.10$0.50$0.0101.09s155 tok/s
98.97%
$0.10$0.50$0.0101.21s110 tok/s
100.00%
Not used in Standard routing:Why these endpoints are not used
$0.11$0.55$0.0110.44s180 tok/s
100.00%
$0.11$0.55$0.0110.51s186 tok/s
99.97%
$0.11$0.55$0.0111.41s141 tok/s
100.00%
$0.11$0.55$0.0111.04s60 tok/s
100.00%
$0.11$0.55$0.0112.42s144 tok/s
100.00%
ProviderInput / Output /M
Input $0.10 / Output $0.50
166 tok/s99.50% uptime
Input $0.10 / Output $0.50
170 tok/s100.00% uptime
Input $0.10 / Output $0.50
167 tok/s100.00% uptime
Input $0.10 / Output $0.50
155 tok/s98.97% uptime
Input $0.10 / Output $0.50
110 tok/s100.00% uptime
Not used in Standard routing:Why these endpoints are not used
Input $0.11 / Output $0.55
180 tok/s100.00% uptime
Input $0.11 / Output $0.55
186 tok/s99.97% uptime
Input $0.11 / Output $0.55
141 tok/s100.00% uptime
Input $0.11 / Output $0.55
60 tok/s100.00% uptime
Input $0.11 / Output $0.55
144 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
$0.08001/M tokens
Output
$0.5704/M tokens
Claude Haiku 5.5 — Price history
Price history
No effective price data for this selection
$0.0567$0.52267.0%9.6%
$0.0745$0.57061.8%43.9%
$0.0852$0.57755.6%17.0%
$0.0879$0.58957.5%9.8%
$0.0979$0.58055.6%16.7%
Not used in Standard routing:Why these endpoints are not used
$0.0678$0.70167.6%1.9%
$0.0760$0.93070.7%0.5%
$0.0845$0.54837.7%0.0%
$0.1061$1.33439.7%0.1%
$0.1116$0.58322.5%0.6%
ProviderInput / Output /M
Token share
Effective in $0.0567 / Effective out $0.522
Token share
9.6%
Effective in $0.0745 / Effective out $0.570
Token share
43.9%
Effective in $0.0852 / Effective out $0.577
Token share
17.0%
Effective in $0.0879 / Effective out $0.589
Token share
9.8%
Effective in $0.0979 / Effective out $0.580
Token share
16.7%
Not used in Standard routing:Why these endpoints are not used
Effective in $0.0678 / Effective out $0.701
Token share
1.9%
Effective in $0.0760 / Effective out $0.930
Token share
0.5%
Effective in $0.0845 / Effective out $0.548
Token share
0.0%
Effective in $0.1061 / Effective out $1.334
Token share
0.1%
Effective in $0.1116 / Effective out $0.583
Token share
0.6%
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
186tok/s
P50, best across providers
Latency
0.44s
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)The model was reachable. Request routed to a provider.
99.99%
Availability (3d)The model returned inference from any provider. Errors and empty responses count against it.
99.81%
Availability over the last 3 days
Last 72 hours
Availability 99.81%
3 Days Ago2 Days AgoYesterdayNow
Availability over the last 24 hours
OpenRouter Availability
99.94%
Without Routing
97.98%
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)
SourceBenchmarkScore
Artificial AnalysisClaude Haiku 5.5 (Xhigh) Intelligence Index34.6
Artificial AnalysisClaude Haiku 5.5 (Xhigh) HLE42.7%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) AA-LCR78.3%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) GDPval-AA51.2%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) CritPt22.6%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) SciCode51.7%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) Terminal-Bench 4.029.3%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) AA-Omniscience Accuracy33.8%
Artificial AnalysisClaude Haiku 5.5 (Xhigh) AA-Omniscience Non-Hallucination Rate60.8%
Artificial AnalysisClaude Haiku 5.5 (Low) Intelligence Index18.6
Artificial AnalysisClaude Haiku 5.5 (Low) HLE27.0%
Artificial AnalysisClaude Haiku 5.5 (Low) AA-LCR72.0%
Artificial AnalysisClaude Haiku 5.5 (Low) GDPval-AA31.5%
Artificial AnalysisClaude Haiku 5.5 (Low) CritPt9.1%
Artificial AnalysisClaude Haiku 5.5 (Low) SciCode49.2%
Artificial AnalysisClaude Haiku 5.5 (Low) Terminal-Bench 4.012.6%
Artificial AnalysisClaude Haiku 5.5 (Low) AA-Omniscience Accuracy30.4%
Artificial AnalysisClaude Haiku 5.5 (Low) AA-Omniscience Non-Hallucination Rate59.5%
Artificial AnalysisClaude Haiku 5.5 (High) Intelligence Index29.9
Artificial AnalysisClaude Haiku 5.5 (High) HLE37.3%
Artificial AnalysisClaude Haiku 5.5 (High) AA-LCR77.3%
Artificial AnalysisClaude Haiku 5.5 (High) GDPval-AA46.2%
Artificial AnalysisClaude Haiku 5.5 (High) CritPt18.6%
Artificial AnalysisClaude Haiku 5.5 (High) SciCode48.7%
Artificial AnalysisClaude Haiku 5.5 (High) Terminal-Bench 4.021.7%
Artificial AnalysisClaude Haiku 5.5 (High) AA-Omniscience Accuracy31.8%
Artificial AnalysisClaude Haiku 5.5 (High) AA-Omniscience Non-Hallucination Rate61.1%
Artificial AnalysisClaude Haiku 5.5 (Medium) Intelligence Index24.6
Artificial AnalysisClaude Haiku 5.5 (Medium) HLE33.8%
Artificial AnalysisClaude Haiku 5.5 (Medium) AA-LCR77.3%
Artificial AnalysisClaude Haiku 5.5 (Medium) GDPval-AA39.1%
Artificial AnalysisClaude Haiku 5.5 (Medium) CritPt12.9%
Artificial AnalysisClaude Haiku 5.5 (Medium) SciCode49.0%
Artificial AnalysisClaude Haiku 5.5 (Medium) Terminal-Bench 4.015.2%
Artificial AnalysisClaude Haiku 5.5 (Medium) AA-Omniscience Accuracy31.7%
Artificial AnalysisClaude Haiku 5.5 (Medium) AA-Omniscience Non-Hallucination Rate59.6%
Artificial AnalysisClaude Haiku 5.5 (Max) Intelligence Index36.5
Artificial AnalysisClaude Haiku 5.5 (Max) HLE44.4%
Artificial AnalysisClaude Haiku 5.5 (Max) AA-LCR82.7%
Artificial AnalysisClaude Haiku 5.5 (Max) GDPval-AA56.4%
Artificial AnalysisClaude Haiku 5.5 (Max) CritPt18.9%
Artificial AnalysisClaude Haiku 5.5 (Max) SciCode55.0%
Artificial AnalysisClaude Haiku 5.5 (Max) Terminal-Bench 4.032.8%
Artificial AnalysisClaude Haiku 5.5 (Max) AA-Omniscience Accuracy34.7%
Artificial AnalysisClaude Haiku 5.5 (Max) AA-Omniscience Non-Hallucination Rate65.8%
Design ArenaClaude Haiku 5.5 Models Arena 3D Elo1322
Design ArenaClaude Haiku 5.5 Models Arena Asciiart Elo1200
Design ArenaClaude Haiku 5.5 Models Arena Code Categories Elo1277
Design ArenaClaude Haiku 5.5 Models Arena Data Visualization Elo1278
Design ArenaClaude Haiku 5.5 Models Arena Game Development Elo1314
Design ArenaClaude Haiku 5.5 Models Arena SVG Elo1195
Design ArenaClaude Haiku 5.5 Models Arena UI Component Elo1272
Design ArenaClaude Haiku 5.5 Models Arena Website Elo1253
Vals AIClaude Haiku 5.5 BioMysteryBench70.4%
Vals AIClaude Haiku 5.5 ClaimBench100.0%
Vals AIClaude Haiku 5.5 Code Migration v1.152.4%
Vals AIClaude Haiku 5.5 CyberBench v1.175.5%
Vals AIClaude Haiku 5.5 EMB v1.165.5%
Vals AIClaude Haiku 5.5 Finance Agent v2.165.4%
Vals AIClaude Haiku 5.5 Vals Finance Index57.4%
Vals AIClaude Haiku 5.5 HLAB-Vals-1.14.2%
Vals AIClaude Haiku 5.5 Immigration61.8%
Vals AIClaude Haiku 5.5 IOI47.3%
Vals AIClaude Haiku 5.5 Vals Legal Index70.3%
Vals AIClaude Haiku 5.5 Legal Research Bench43.8%
Vals AIClaude Haiku 5.5 MedCode51.5%
Vals AIClaude Haiku 5.5 MedScribe89.5%
Vals AIClaude Haiku 5.5 MysteryMechanism35.1%
Vals AIClaude Haiku 5.5 ProgramBench2.5%
Vals AIClaude Haiku 5.5 ProofBench v1.189.0%
Vals AIClaude Haiku 5.5 Public Benefits Bench v1.163.2%
Vals AIClaude Haiku 5.5 SAGE89.8%
Vals AIClaude Haiku 5.5 Vals Science Index38.5%
Vals AIClaude Haiku 5.5 SRE Bench5.7%
Vals AIClaude Haiku 5.5 SWE-bench Multimodal44.0%
Vals AIClaude Haiku 5.5 Tax Agent Bench31.9%
Vals AIClaude Haiku 5.5 Terminal-Bench 4.036.4%
Vals AIClaude Haiku 5.5 Terminal-Bench-Science10.0%
Vals AIClaude Haiku 5.5 Vals Index58.7%
Vals AIClaude Haiku 5.5 Vibe Code Bench 1-10020.4%
Vals AIClaude Haiku 5.5 Vibe Code Bench v1.190.4%
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.
1.
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.
786Btokens
2.
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.
440Btokens
3.
pi
There are many coding agents, but this one is yours.
144Btokens
4.
payready-document-extraction
new
106Btokens
5.
Freebuff
Powerful coding models, funded by ads
105Btokens
Rest of OpenRouter
208B
Top 5 Apps
34.9B
Activity
Token volume and request traffic to this model over time.
API
Drop-in code to call this model. OpenRouter's API is OpenAI-compatible — most SDKs work by just swapping the base URL. The only thing that changes between models is the model slug below.
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 EuropeGoogle Vertex, Amazon BedrockPrices 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 URLhttps://eu.openrouter.ai/api/v1PlanRequires 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
Documentation
Guides, cookbooks, blog posts, and tools for building with Anthropic: Claude Haiku 5.5 on OpenRouter.
Claude 5.5 Migration GuideCookbook
Reranking in RAG: How to Fix Bad Retrieval with a RerankerBlog post
How to Benchmark LLMsBlog post
How to Stop an AI Agent Stuck in an Infinite LoopBlog post
MCP vs CLI vs Skills: What Each Costs in ContextBlog post
LLM API Rate Limits: How to Fix OpenAI and Claude 429 ErrorsBlog post
How to Get Structured Output from Any LLM: JSON Schema, Tool Calls, or a Decision ModelBlog post
Best OCR Model for Scanned Documents, MeasuredBlog post
Build an AI Code Reviewer with OpenRouterBlog post
How to Build Your Own Always-On AI AgentBlog post
Claude Opus vs Sonnet vs Haiku: Which Should You Use?Blog post
How Much Does Claude Code Cost? Real Session Costs by ModelBlog post
How to Extract Tables from a PDF with an LLMBlog post
Hit Your Claude Code Usage Limit? How to Keep CodingBlog post
Claude 529 Overloaded: What It Means and How to Fix ItBlog post
Frequently asked questions
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.
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.
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.
Claude Haiku 5.5 has a 1,000,000 token context window. It supports up to 128,000 completion tokens.
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.
Claude Haiku 5.5 accepts text, images, and files such as PDFs as input and returns text.
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.
Claude Haiku 5.5 was released on October 7, 2026.
正文以来源语言显示。