Anthropic · Claude · Flagship
Claude Opus 5.5
Anthropic's most capable model for long agentic coding sessions, deep refactors and hard reasoning.
Price per 1M tokens
- Input
- $4.40
- Output
- $22.00
- Cached input
- $0.2295% below input
Output is 5× the input rate. $1 covers about 64 prompts of 1,000 input + 500 output tokens. 1 credit = $1 of usage. All rates
Ways to pay
- Pay as you goPrepaid balance at the rates above, top-ups from $3
- Pro $5/monthWeekly allowance of about 185.5k uncached input tokens of this model
- Plus $9/monthWeekly allowance of about 349.1k uncached input tokens of this model, 2% below the rates above
- Max $29/monthWeekly allowance of about 1.16M uncached input tokens of this model, 4% below the rates above
Plans are opening soon; until then every model is pay as you go. How plans will work
Start using Claude Opus 5.5
Set RELAYKIT_API_KEY to a key from your console, then run any example. Streaming is on so long answers are not cut off by idle-connection timeouts.
curl https://api.oneaiapikey.com/v1/chat/completions \
-H "Authorization: Bearer $RELAYKIT_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "claude-opus-5-5",
"stream": true,
"messages": [{"role": "user", "content": "In two sentences, when should I use agentic coding?"}]
}'import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.oneaiapikey.com/v1",
api_key=os.environ["RELAYKIT_API_KEY"],
)
stream = client.chat.completions.create(
model="claude-opus-5-5",
messages=[{"role": "user", "content": "In two sentences, when should I use agentic coding?"}],
stream=True, # recommended for long outputs
)
for chunk in stream:
if chunk.choices and chunk.choices[0].delta.content:
print(chunk.choices[0].delta.content, end="", flush=True)import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.oneaiapikey.com/v1",
apiKey: process.env.RELAYKIT_API_KEY,
});
const stream = await client.chat.completions.create({
model: "claude-opus-5-5",
messages: [{ role: "user", content: "In two sentences, when should I use agentic coding?" }],
stream: true,
});
for await (const chunk of stream) {
process.stdout.write(chunk.choices[0]?.delta?.content ?? "");
}import os
from anthropic import Anthropic
client = Anthropic(
base_url="https://api.oneaiapikey.com",
api_key=os.environ["RELAYKIT_API_KEY"],
)
with client.messages.stream(
model="claude-opus-5-5",
max_tokens=1024,
messages=[{"role": "user", "content": "In two sentences, when should I use agentic coding?"}],
) as stream:
for text in stream.text_stream:
print(text, end="", flush=True)# macOS / Linux
export ANTHROPIC_BASE_URL="https://api.oneaiapikey.com"
export ANTHROPIC_AUTH_TOKEN="$RELAYKIT_API_KEY"
export ANTHROPIC_MODEL="claude-opus-5-5"
claude
# Windows PowerShell
$env:ANTHROPIC_BASE_URL = "https://api.oneaiapikey.com"
$env:ANTHROPIC_AUTH_TOKEN = $env:RELAYKIT_API_KEY
$env:ANTHROPIC_MODEL = "claude-opus-5-5"
claudeMore: quickstart · API reference · try it in Chat first
Overview
When Claude Opus 5.5 fits
Claude Opus 5.5 is a flagship model from Anthropic, available here under the exact model id “claude-opus-5-5”. It sits at the top of the Claude line-up, so it is the one to reach for when answer quality matters more than cost per token.
It costs $4.40 per million input tokens and $22 per million output tokens, so output is 5× the input rate. Long answers are therefore the main cost driver: cap max tokens, ask for diffs instead of whole files, and keep reasoning effort proportional to the task. With a 3:1 input-to-output mix it ranks 10th lowest-cost of the 11 models on this site.
Cached input bills at $0.22 per million tokens, 95% below the standard input rate. Workloads that resend the same system prompt, tool definitions and files on every turn, coding agents in particular, get the most from this.
You can call it through the OpenAI format at https://api.oneaiapikey.com/v1 or the native Anthropic Messages format at https://api.oneaiapikey.com/v1/messages. That makes it usable in Claude Code as well as OpenAI-compatible tools such as Codex CLI, Cursor and Cline.
Cost examples
What a workload costs
Worked from the published rates, with the arithmetic shown. Real numbers depend on prompt size and how much the model writes.
1M input + 200k output tokens
1 × $4.40 + 0.2 × $22.00
$8.80
1,000 chat replies (1.5k in, 400 out each)
1000 × (0.0015 × $4.40 + 0.0004 × $22.00)
$15.40
About 64 replies per credit
One agent turn (30k in, 24k of it cached, 2k out)
0.006 × $4.40 + 0.024 × $0.22 + 0.002 × $22.00
$0.07568
$0.176 with no cache hits. Writing tokens into the cache bills 1.25× the input rate (2× for a 1-hour cache) when the provider reports cache writes
Works with
Compatible agents 8
Coding tools that speak the Anthropic or OpenAI format. Each guide sets the base URL, key and this model id.
Similar models
Same tier, other vendors
Same key, same balance. Switching is a change to the model field.
Try Claude Opus 5.5 from a $3 top-up.
There is no sign-up credit: add $3 or more to your wallet and every request is billed at the rates above, with a receipt in your usage log.
Claude Opus 5.5 is developed by Anthropic. Model names and trademarks belong to their owners. Relaykit is an independent API gateway and is not affiliated with, endorsed by or sponsored by Anthropic.