ai.canshu.com
中文
VRAM Calculator Advisor & TCO API Pricing Compare Models Leaderboard GPUs

Search models, GPUs, and tools

↑↓ Navigate · Enter Open · ★ Watchlist prioritized ai.canshu.com
Anthropic
Proprietary API Anthropic · claude-haiku-5-5 · Text+Image+PDF

Claude Haiku 5.5 Official Specifications, Pricing & Benchmarks

Ultra-fast subagent & code-search model from Anthropic at $0.10/$0.50 with 1M context.

📋 Claude Haiku 5.5 Official API & System Card Specifications

Verified against official Anthropic developer documentation & pricing

API Model ID: claude-haiku-5-5

1. Identity & Capabilities

API Spec
API Model ID
claude-haiku-5-5
Release Date
2026-10-07
Weight License
Proprietary (Undisclosed)
Native Reasoning
✅ Supported
Function Calling & JSON Schema
✅ 原生支持
Batch API 异步批处理
✅ Supported (50% off)

2. Context, Output & Modalities

Limits
Max Context Window
1,000,000 tokens (1000K)
Max Output Tokens
128,000 tokens
Input Modalities
Text + Image + PDF
Output Modalities
Text
Direct China Access
🌐 Global / OpenRouter
Deployment Mode
Cloud API only (Proprietary)

3. Official Pricing & Cache

USD / 1M
Input / 1M Tokens
$ 0.1
Cached Input / 1M
$ 0.01 -90%
Output / 1M Tokens
$ 0.5
Tiered Pricing Notes
$0.10/$0.50 ($0.01 cached) with full 1M context and Extended Thinking support

🏆 Official Benchmark Results

Sourced from official System Card / Model Card & technical reports

Arena Elo ~1428 AA Index 41
💻 Coding
SWE-bench Verified 68.5 %

A verified subset of 500 software engineering problems from real GitHub issues, validated by human annotators for evaluating language models' ability to resolve real-world coding issues by generating patches for Python codebases.

📚 Knowledge
MMLU-Pro 79.2 %

A more robust and challenging multi-task language understanding benchmark that extends MMLU by expanding multiple-choice options from 4 to 10, eliminating trivial questions, and focusing on reasoning-intensive tasks.

🧠 Reasoning
GPQA Diamond 73.5 %

PhD-level Google-proof science reasoning questions

🤖 Agentic
BFCL-V4 69 %

Berkeley Function Calling Leaderboard V4 (BFCL-V4) evaluates LLMs on their ability to accurately call functions and APIs, including simple, multiple, parallel, and nested function calls across diverse programming scenarios.

🎛️ Official Sampling & Reasoning Parameters

Recommended generation settings & API parameter constraints

Fast Routing & Subagent Mode
temperature=0.3 top_p=0.95 max_tokens=16,384

📊 Claude Haiku 5.5 Full Benchmark Scorecard

34 public benchmarks · ranks among 154 models · synced 2026-11-20

⚔️ Compare head-to-head

💻 Coding 11 items

BenchCAD (with Python tool) #5/8 87%

BenchCAD variant evaluated with access to a Python tool for programmatic CAD reasoning.

SWE-bench Multilingual #5/33 83.7%

A multilingual benchmark for issue resolving in software engineering that covers Java, TypeScript, JavaScript, Go, Rust, C, and C++.

Program Bench #3/18 82%

Program Bench evaluates code-generation agents by asking them to recreate a program's behavior from only a compiled binary and documentation.

BenchCAD #5/8 67%

BenchCAD is a benchmark for programmatic CAD reasoning built from 17,900 execution-verified CadQuery programs spanning 106 industrial part families, roughly half anchored to real ISO, DIN, EN, ASME, and IEC specification tables.

SWE-Bench Pro #7/52 64.8%

SWE-Bench Pro is an advanced version of SWE-Bench that evaluates language models on complex, real-world software engineering tasks requiring extended reasoning and multi-step problem solving.

FrontierCode 1.1 (Extended) #3/3 58.4%

FrontierCode 1.1 Extended subset: agentic coding tasks graded for mergeable patches, reported separately from the Main subset.

FrontierCode 1.1 #10/23 46.4%

FrontierCode 1.1 evaluates whether coding-agent changes are mergeable, using unit tests, maintainer-defined rubrics, and verifiers. Runs flagged for unfair internet use receive a zero score.

FrontierSWE V2 #6/7 43.8%

FrontierSWE V2 evaluates agents on open-ended technical projects and reports mean@5. Kept separate from FrontierSWE V1 dominance-score results, which are not comparable.

Terminal-Bench 4.0 #16/31 39.2%

Terminal-Bench 4.0 evaluates AI agents on real-world work in terminal and command-line containerized environments across 66 tasks, with emphasis on science-adjacent and frontier engineering problems.

SWE-Bench Multimodal #6/7 30.7%

SWE-Bench Multimodal extends SWE-Bench to evaluate language models on software engineering tasks that involve visual inputs such as screenshots, UI mockups, and diagrams alongside code understanding.

Terminal-Bench-Science 0.1 #16/16 20.6%

Terminal-Bench-Science 0.1 is a Stanford-led community benchmark of 70 tasks drawn from scientific research workflows across the life, physical, earth, mathematical, and engineering sciences.

🤖 Agents & Tools 9 items

OfficeQA #4/4 73.5%

OfficeQA is a Databricks public benchmark that evaluates end-to-end grounded reasoning over a large corpus of historical U.S. Treasury Bulletin documents. Models must locate relevant tables across the corpus and perform precise numerical reasoning over them.

OSWorld 2.1 (partial) #6/7 72.4%

Partial-credit computer-use score on OSWorld 2.1, as reported in the Claude Sonnet 5.5 release. Separate from OSWorld 2.0 and strict completion scores.

LatchBio SpatialBench Verified #3/3 67.7%

LatchBio SpatialBench Verified tests analysis of spatial transcriptomics data across externally validated problems, requiring the model to answer biological questions about the sample from analysis results.

OfficeQA Pro #11/15 60.3%

OfficeQA Pro evaluates AI models on professional knowledge-work questions and tasks drawn from real office workflows, including document analysis, spreadsheet reasoning, and information synthesis across business domains.

LatchBio SingleCellBench #3/3 56.2%

LatchBio SingleCellBench tests analysis of single-cell RNA sequencing data across problems spanning standard workflows such as labeling cell types, finding differentially expressed genes, and correcting batch effects.

GDPval-AA 2.1 #6/7 1620

Version 2.1 of the Artificial Analysis GDPval evaluation of professional knowledge work, reported as an Elo rating.

AA-Briefcase v1.1 #6/7 1578

AA-Briefcase v1.1 evaluates multi-hour professional office work and reports Elo scores. Kept separate from unversioned AA-Briefcase results.

PhysicianBench #3/3 43%

PhysicianBench is a public benchmark of 100 physician tasks carried out in an electronic health record (EHR), measuring agentic clinical task completion.

OSWorld 2.1 (strict) #3/4 37.1%

Strict pass rate on OSWorld 2.1: fraction of tasks where every graded checkpoint is satisfied. Separate from the OSWorld 2.1 partial-credit score and from OSWorld 2.0.

Source: ZeroEval / llm-stats.com (✓ = third-party verified, others vendor-reported).

Closest Self-Hostable Open-Weight Alternatives to Claude Haiku 5.5

Open-weight models with full HuggingFace weights for private GPU / Mac deployment

Compare with these →

Frequently Asked Questions

Is Claude Haiku 5.5 open-source or proprietary?

Claude Haiku 5.5 (claude-haiku-5-5) is a proprietary commercial model by Anthropic. Weights and internal layer dimensions are undisclosed; it is accessed via cloud API.

What are the context limit and API pricing of Claude Haiku 5.5?

Claude Haiku 5.5 supports 1000K input context and 128,000 max output tokens, priced at $0.1/1M input ($0.01/1M cached) and $0.5/1M output.