AnthropicMediumUltra

Claude Opus 5.5

Anthropic's current Opus model (September 2026), cheaper than Opus 5 at $4/$20 per M tokens. 1M-token context window and 128K max output tokens, with prompt caching (read $0.20/M, i.e. 5% of input; 5-minute write $5.00/M) and web search. Compatible with the Anthropic Messages and OpenAI Chat Completions formats.

7,432/37,160credits

input / output · per 1M tokens

Current Opus model

1M token context window

128K max output tokens

Prompt caching (read 5% / 5-min write 125% of input price)

Web search tool support

Vision and tool use

Streaming support

Run it right now

Test this model instantly in the Console Playground — no code required

Use with AI Assistant

Copy usage instructions for Claude, ChatGPT, or other AI

llms.txtllms.txt ↗

Model Specifications

Context Window

tokens

Max Output

128K

tokens

Training Cutoff

Not published

Compatible SDK

Anthropic, OpenAI

Capabilities

Vision

Function Calling

Streaming

JSON Mode

System Prompt

Token Pricing (per 1M tokens)

Token Type	Credits	USD Equivalent
Input Tokens	7,432	$4.95
Output Tokens	37,160	$24.77
Cached Tokens	371.6	$0.25

* 1,500 credits ≈ $1 (actual charges may vary based on usage)

Quick Start

curl -X POST "https://api.core.today/llm/anthropic/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer cdt_your_api_key" \
  -d '{
  "model": "claude-opus-5-5",
  "messages": [
    {
      "role": "system",
      "content": "You are a principal research analyst. Provide rigorous, source-aware analysis."
    },
    {
      "role": "user",
      "content": "Given this corpus of quarterly reports, identify the three biggest emerging risks and quantify their exposure."
    }
  ],
  "max_tokens": 8000
}'

Parameters

Parameter	Type	Required	Default	Description
`messages`	array	Yes	-	Array of message objects (OpenAI format)
`max_tokens`	integer	No	4096	Maximum tokens in response (up to 128000)
`stream`	boolean	No	false	Enable Server-Sent Events streaming

Examples

Long-context Analysis

Deep multi-step analysis over large documents with Claude Opus 5.5

curl -X POST "https://api.core.today/llm/anthropic/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer cdt_your_api_key" \
  -d '{
  "model": "claude-opus-5-5",
  "messages": [
    {
      "role": "system",
      "content": "You are a principal research analyst. Provide rigorous, source-aware analysis."
    },
    {
      "role": "user",
      "content": "Given this corpus of quarterly reports, identify the three biggest emerging risks and quantify their exposure."
    }
  ],
  "max_tokens": 8000
}'

Production Code Generation

Generate production-ready code with Claude Opus 5.5

curl -X POST "https://api.core.today/llm/anthropic/v1/messages" \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer cdt_your_api_key" \
  -d '{
  "model": "claude-opus-5-5",
  "messages": [
    {
      "role": "system",
      "content": "You are a senior software engineer. Write clean, well-tested, production code."
    },
    {
      "role": "user",
      "content": "Implement an idempotent payment reconciliation worker in Python with retries and unit tests."
    }
  ],
  "max_tokens": 6000
}'

Tips & Best Practices

1Cheaper than Opus 5 ($4/$20 vs $5/$25) — prefer it over Opus 5 for new workloads; escalate to Claude Fable 5.1 only for the hardest tasks

2Prompt caching: read $0.20/M (5%), 5-minute write $5.00/M — cache large repeated contexts

31M context window handles very large codebases and document sets in a single request

4Web search costs about 18.6 credits per request ($10 per 1,000 searches) on top of token usage

5Sampling params (`temperature`, `top_p`, `top_k`) are not accepted by Claude 4.7+ models — omit them (the console strips `temperature` automatically)

6Streaming recommended for long responses

7Compatible with both Anthropic Messages and OpenAI SDK formats

8Fast mode (speed: fast) bills token charges at 2x, and US-only inference (inference_geo: us) at 1.1x