BETA

Claude Opus 5.5 pricing updated.

Model prices Latest updates in Telegram
View prices Contact sales Telegram support
Help
Support Channel Feedback Telegram
EN
EN English RU Русский 中文 中文 한국어 한국어 TR Türkçe
Log in Try
Web Chat · API · Hermes

Top models and agents. One workspace.

Claude, GPT, and Fable in Web Chat, Telegram, IDEs, and your own API. Managed Hermes keeps working while you are away.

GPT 6.1 Sol in Web Chat, API and Hermes — same rates as GPT 6 Sol.

Try ApiClo See pricing
$1.24
minimum top-up
From 4¢
input / 1M tokens
Pay only
for usage
Prompt cache: see read and write rates by model →
Web ChatDelivered requests, saved history
Claude Fable 5.1
You

Explain function calling in simple terms.

AI

Function calling lets a model send structured data to an external tool or API, then use the result in its answer.

Message a model…
Delivered
Connect where you work
01 / Connect

Meet ApiClo where you work.

Connect Telegram or your IDE and keep the same models, balance, and account.

Your IDE

Ready to connect
Cursor · VS Code · JetBrainsSetup guides
Use one API
02 / API

One endpoint. Your model.

Use an OpenAI-compatible API with the same account balance and selected-model pricing.

Read the docs →View pricing →
POST/v1/chat/completions
base_url: https://apiclo.com/v1
{
  "model": "claude-fable-5-1",
  "messages": [
    {"role": "user", "content": "Explain function calling."}
  ],
  "stream": false
}
200 Delivered

Function calling lets a model send structured data to an external tool or API.

usage: prompt 23 · completion 18 · total 41

Check the models

See the checks. Try your own.

Open the original reports or repeat the checks with your own ApiClo key.

Your key. The same check.

Create an API key in your account. Check Opus with CCTest or Hvoy, and Astra with Hvoy. Use the endpoint and model ID below; paste your key directly on the test site.

API endpoint
https://apiclo.com
Claude Opus 5.5
claude-opus-5-5
GPT 6 Astra
gpt-6-astra
Create a key CCTest

Each report records one test run. Use your own key to check the current result. Test requests are billed at the selected model's rate.

03 / Hermes

Work continues after you leave.

Managed Hermes agents run tasks in dedicated workspaces. History and status always stay in view.

Open Hermes
Summarize product updatesRunning
workspace · hermes-prod
  1. Task received10:25:02
  2. Gathering contextRecent updates and related documents
  3. Writing summary
  4. Complete
Live result

Product update brief

Platform stability, new API capabilities, and developer experience improvements are being combined into one cited brief.

product-updates.md
Dedicated workspacePersistent historyStatus in view

Managed Hermes hosting is billed separately. Included Hermes tokens are separate from the cash balance and work only inside managed Hermes.

04 / Pricing

Pay for what you use.

No mandatory API subscription. Input, output, and prompt-cache rates follow your selected model.

$1.24
minimum top-up
From 4¢
input / 1M tokens
One account
Web Chat · API · Hermes
Selected-model rate is shown before every request

Example cost with prompt caching per 1M total tokens: GPT 6 Luna — $0.026 with cache; without cache: $0.040. Scenario: 40% regular input, 40% cache-read input, 20% output. Cache-write costs are extra; this is an example, not a customer average. See the formula and all models.

All model rates, including cache
ModelInput / 1MCache read / 1MCache write / 1MOutput / 1M
Claude Opus 5.5$0.40$0.06 · 0.15×$0.50 · 5 min
$0.80 · 1 h$0.40
GPT 6 Sol$0.40$0.04 · 0.10×$0.50$0.40
GPT 6 Luna$0.04$0.004 · 0.10×$0.05$0.04
GPT 6 Astra$0.60$0.12 · 0.20×$0.75$0.60
Claude Fable 5.1$1.20$0.16 · 0.13×$1.50 · 5 min
$2.40 · 1 h$1.20
GPT 5.6 Sol$0.40$0.04 · 0.10×$0.50$0.40
Claude Opus 5$0.40$0.04 · 0.10×$0.50 · 5 min
$0.80 · 1 h$0.40
Claude Sonnet 5.5$0.35$0.035 · 0.10×$0.4375 · 5 min
$0.70 · 1 h$0.35
Claude Sonnet 5$0.38$0.038 · 0.10×$0.475 · 5 min
$0.76 · 1 h$0.38
GPT 5.6 Terra$0.30$0.03 · 0.10×$0.375$0.30
GPT 5.6 Luna$0.19$0.019 · 0.10×$0.238$0.19
Claude Fable 5$0.80$0.08 · 0.10×$1.00 · 5 min
$1.60 · 1 h$0.80
Claude Opus 4.8$0.40$0.04 · 0.10×$0.50 · 5 min
$0.80 · 1 h$0.40
Claude Opus 4.7$0.29$0.029 · 0.10×$0.363 · 5 min
$0.58 · 1 h$0.29
Claude Sonnet 4.6$0.32$0.032 · 0.10×$0.40 · 5 min
$0.64 · 1 h$0.32
Claude Haiku 4.5$0.32$0.032 · 0.10×$0.40 · 5 min
$0.64 · 1 h$0.32
Claude Opus 4.6$0.29$0.029 · 0.10×$0.363 · 5 min
$0.58 · 1 h$0.29
Claude Sonnet 4.5$0.32$0.032 · 0.10×$0.40 · 5 min
$0.64 · 1 h$0.32
Claude Opus 4.5$0.40$0.04 · 0.10×$0.50 · 5 min
$0.80 · 1 h$0.40
GPT 6.1 Sol$0.40$0.04 · 0.10×$0.50$0.40

All rates are USD per 1M tokens. Cache rates apply only when the provider reports cache-read or cache-write tokens. See the cache-pricing guide.

Your selected model is tried first. If a documented provider or capacity failure triggers a controlled fallback, billing keeps the selected-model tariff. Managed Hermes hosting is billed separately.

05 / FAQ

Questions, answered.

Do I need a subscription?

No mandatory plan for API, Web Chat, or Telegram: add balance and pay for actual token usage. Managed Hermes hosting has a separate monthly charge.

Where can I use ApiClo?

In Web Chat, Telegram, supported IDEs, OpenAI-compatible clients, and your own backend.

What is Hermes?

A managed agent workspace with files, plugins, persistent history, and visible run status for multi-step work.

One account. Every way to work.

Start with Web Chat, connect your tools, or launch Hermes.

Try ApiCloView pricing