Home/Models/Anthropic/Claude Opus 4.5
C

Claude Opus 4.5

Input:$4/M
Output:$20/M
Context:200K
Max Output:64K
Claude Opus 4.5 is an instruction-tuned large language model from Anthropic designed for complex reasoning, coding, and multi-turn dialogue. It supports extended context handling, tool/function calling, structured outputs, and integration with retrieval-augmented workflows. Typical uses include analytical assistants, code generation and review, knowledge-base QA, and content drafting with policy-aligned responses. Technical highlights include instruction following, RAG-friendly behavior, and safety controls available in Claude deployments.
New
Commercial Use
Playground
Overview
Features
Pricing
API
Versions

What is Claude Opus 4.5

Claude Opus 4.5 is Anthropic’s newest “Opus”-class large language model (LLM) in the Claude 4.5 family, built for complex reasoning, professional software engineering, and long-running agentic workflows. Anthropic positions Opus 4.5 as a top-tier option that balances maximum capability with practical performance and a more accessible price point than prior Opus releases. The release completes the Claude 4.5 family alongside Sonnet 4.5 and Haiku 4.5.

Model family / identifier: Claude Opus 4.5 (API name noted by Anthropic as claude-opus-4-5-20251101).


Core features of Claude Opus 4.5

  • Purpose / primary strengths: Professional software engineering, advanced agents, spreadsheet/financial modeling, extended multi-turn workflows, and “computer use” (interacting with GUIs, spreadsheets, browsers).
  • Effort parameter: New request parameter (low / medium / high) that trades off compute/tokens for depth of reasoning. Useful for tuning latency/cost vs answer thoroughness.
  • Enhanced computer use & vision: New zoom action and improved screen/UI inspection so the model can read fine UI elements and detailed visuals before taking actions.
  • Thinking continuity: “Thinking block preservation” — Opus 4.5 preserves prior reasoning blocks to maintain continuity across long sessions.

Technical details of Claude Opus 4.5

Model identifier: claude-opus-4-5-20251101 (commonly referenced for API calls).

Context window & token limits: input 200,000 tokens / output 64,000 tokens

Modalities: text, code, images and document inputs supported; outputs are text (structured and unstructured). Partner integrations add features (batch predictions, function calling, prompt caching).

Benchmark performance & empirical results

Anthropic and multiple independent outlets have published/reported benchmark results and test behavior for Opus 4.5. Below are the most notable claims and figures:

  • SWE-bench (software engineering benchmarks) — SWE-bench ≈ 80.9% in Anthropic’s reported evaluations. Opus 4.5 leads in 7/8 programming languages on Anthropic’s SWE-bench Multilingual tests.
  • Real-world agent endurance — improved continuous agent operation and better handling of multi-step workflows and tool orchestration.
  • Aider Polyglot: +10.6% improvement vs Sonnet 4.5 on complex coding tasks. Anthropic
  • Terminal Bench: +15% improvement over Sonnet 4.5 in multi-step terminal/code workflows.
  • Vending-Bench (long-horizon planning): 29% improvement over Sonnet 4.5 on long-horizon agentic tasks.
  • Autonomous coding: Anthropic reports consistent performance through 30-minute autonomous coding sessions (versus earlier models that decayed earlier).

Typical and priority use cases

  • Complex software engineering & long scripts — code generation, debugging, multi-file refactors, and autonomous coding agents.
  • Autonomous agents and tool orchestration — chaining API calls, browsing, spreadsheet automation, and long multi-step workflows where state must be preserved across many turns.
  • Large document synthesis & research — legal briefs, long reports, multi-chapter writing, and summarization over very large corpora thanks to extended context and compaction.
  • Enterprise automation — internal tooling, data extraction from files/spreadsheets, and agents that operate business processes continuously.

Claude Opus 4.5 vs Gemini 3.0 Pro vs GPT 5.1

DimensionClaude Opus 4.5 (Anthropic)Gemini 3.0 Pro (Google)GPT 5.1 (OpenAI)
Model nameclaude-opus-4-5-20251101-thinking;claude-opus-4-5-20251101gemini-3-pro-preview-thinking;gemini-3-pro-previewgpt-5.1-chat-latest;gpt-5.1
Primary strengthsLong-horizon agentic reliability, token efficiency, coding & multi-agent orchestration, safety focus.Frontier reasoning, multimodal performance, top leaderboard scores (LMArena, GPQA, MathArena); broad Google integration.Adaptive reasoning, developer ergonomics, extended prompt caching (24h), fast interactivity and coding tooling.
Representative pricing (input/output per 1M tokens)$5 / $25 (Anthropic stated Opus 4.5). — enterprise preview pricing.Reported preview tiers ≈ $2 / $12 or higher tiers (varies by region / plan).$1.25 / $10 (OpenAI published GPT-5.1 API pricing).
Context / window & memoryFocus on context compaction, memory features, and efficient long sessions; tuned for multi-agent runs.Very large context windows reported (1M tokens in preview) and multimodal inputs (text, image, audio, video).Extended prompt caching and efficient adaptive reasoning; caching intended to reduce cost & latency for follow-ups.
Best fit forEnterprise agents, long autonomous code workflows, cost-sensitive long runs, regulated environments with safety controls.Research/benchmarks, multimodal reasoning, integrated Google ecosystem features (Search/Ads/Apps).Interactive developer products, low-latency coding assistance, iterative workflows benefiting from prompt caching.

How to access Claude opus 4.5 API

Step 1: Sign Up for API Key

Log in to cometapi.com. If you are not our user yet, please register first. Sign into your CometAPI console. Get the access credential API key of the interface. Click “Add Token” at the API token in the personal center, get the token key: sk-xxxxx and submit.

Step 2: Send Requests to Claude opus 4.5 API

Select the “claude-opus-4-5-20251101-thinking;claude-opus-4-5-20251101” endpoint to send the API request and set the request body. The request method and request body are obtained from our website API doc. Our website also provides Apifox test for your convenience. Replace <YOUR_API_KEY> with your actual CometAPI key from your account. base url is Anthropic Messages format and Chat format.

Insert your question or request into the content field—this is what the model will respond to . Process the API response to get the generated answer.

Step 3: Retrieve and Verify Results

Process the API response to get the generated answer. After processing, the API responds with the task status and output data.

See also Gemini 3 Pro Preview API

FAQ

Can Claude Opus 4.5 complete multi-day software projects in hours?

Yes, Claude Opus 4.5 can confidently deliver multi-day software development projects in hours, working independently with technical depth to create efficient solutions. It can even pass Anthropic's most challenging coding interviews.

What benchmark scores does Claude Opus 4.5 achieve for coding tasks?

Claude Opus 4.5 achieves 80.9% on SWE-bench Verified, setting industry-leading results for real-world software coding. It also reaches 66.3% on OSWorld for computer use tasks.

How does the effort parameter work in Claude Opus 4.5?

Opus 4.5 offers hybrid reasoning with configurable effort levels. API users can adjust overall effort to balance performance with latency and cost—lower effort delivers quality efficiently, while higher effort maximizes accuracy for complex tasks.

When should I use Claude Opus 4.5 instead of Claude Sonnet 4.5?

Choose Opus 4.5 for mission-critical tasks requiring frontier intelligence—particularly production-ready code, sophisticated AI agents, and complex document creation where maximum accuracy justifies higher cost.

Does Claude Opus 4.5 support tool calling and function use?

Yes, Opus 4.5 paired with Anthropic's tool use improvements enables more capable agents with complex, multi-step tasks requiring sustained reasoning and adaptive decision-making.

Can Claude Opus 4.5 create spreadsheets, slides, and documents?

Yes, Opus 4.5 delivers a step-change improvement in creating spreadsheets, slides, and docs. It better leverages memory to maintain context and consistency across files for enterprise workflows.

What makes Claude Opus 4.5 different from previous Opus versions?

Opus 4.5 (Nov 2025) represents a meaningful step forward with state-of-the-art coding, improved planning and architecture choices, and the ability to handle sprawling professional projects from start to finish.

Is Claude Opus 4.5 available on cloud platforms like AWS and Google Cloud?

Yes, Opus 4.5 is available on the Claude Developer Platform natively, Amazon Bedrock, Google Cloud's Vertex AI, and Microsoft Foundry.

Features for Claude Opus 4.5

Explore the key features of Claude Opus 4.5, designed to enhance performance and usability. Discover how these capabilities can benefit your projects and improve user experience.

Pricing for Claude Opus 4.5

Explore competitive pricing for Claude Opus 4.5, designed to fit various budgets and usage needs. Our flexible plans ensure you only pay for what you use, making it easy to scale as your requirements grow. Discover how Claude Opus 4.5 can enhance your projects while keeping costs manageable.
Comet Price (USD / M Tokens)Official Price (USD / M Tokens)Discount
Input:$4/M
Output:$20/M
Input:$5/M
Output:$25/M
-20%

Sample code and API for Claude Opus 4.5

Access comprehensive sample code and API resources for Claude Opus 4.5 to streamline your integration process. Our detailed documentation provides step-by-step guidance, helping you leverage the full potential of Claude Opus 4.5 in your projects.
Python
JavaScript
Curl
import anthropic
import os

# Get your CometAPI key from https://api.cometapi.com/console/token, and paste it here
COMETAPI_KEY = os.environ.get("COMETAPI_KEY") or "<YOUR_COMETAPI_KEY>"
BASE_URL = "https://api.cometapi.com"

message = anthropic.Anthropic(
    base_url=BASE_URL,
    api_key=COMETAPI_KEY,
)
messages = message.messages.create(
    model="claude-opus-4-5-20251101",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello, Claude"}],
)
print(messages.content[0].text)

Versions of Claude Opus 4.5

The reason Claude Opus 4.5 has multiple snapshots may include potential factors such as variations in output after updates requiring older snapshots for consistency, providing developers a transition period for adaptation and migration, and different snapshots corresponding to global or regional endpoints to optimize user experience. For detailed differences between versions, please refer to the official documentation.
claude-opus-4-5-20251101-thinkingThere is a ThinkChain version, API usage.
claude-opus-4-5-20251101There is a NonthinkChain version, API usage.
cometapi-opus-4-5-20251101-thinkingThere is a ThinkChain version, cursor usage.
cometapi-opus-4-5-20251101There is a NonThinkChain version, cursor usage.

More Models