Claude API Pricing 2026 — Per-Model Rates, Payment Methods, and Actual Costs

Claude is the standard for reasoning, coding, and agents — and Kunavo offers every Claude model 30% below Anthropic's official rates, depending on the model, behind one OpenAI-compatible API. Since September 24, 2026, this also applies to claude-sonnet-5. Anthropic made the $2.00 / $10.00 rates the permanent standard price on August 31, 2026, and Kunavo sells it at 30% below that rate. This guide covers what Indonesian developers most often search for about Claude API pricing and payment methods: the latest rates by model, how to pay from Indonesia, cost examples you can verify yourself, and the most effective ways to reduce your bill.

Claude API pricing at a glance

Rates in USD per 1M tokens, exactly as billed by Kunavo. The “Anthropic rate” column shows the official price for the same model so you can see the difference at a glance.

Live rates are always available on the pricing page and on each model page (claude-sonnet-4-6, claude-haiku-4-5, claude-opus-4-7). Haiku is for low-cost workloads, Sonnet is a balanced default, and Opus and Fable are for the most demanding reasoning. Claude Opus 5.5, released by Anthropic on September 22, 2026, is the latest Opus: it has a 1M-token context window and supports up to 128K output tokens.

Claude Opus 5.5 offers the best price-to-performance. Anthropic lists it at $4.00 / $20.00 per 1M tokens — below Opus 5 at $5.00 / $25.00 — and says its performance is on par with Claude Fable 5.1 for most workloads. Kunavo's rate is $2.80 / $14.00, about 30% below Anthropic's rate and cheaper than Opus 5 ($3.50 / $17.50); this is Kunavo's lowest-priced Opus rate. Sonnet 5 is even cheaper: $1.40 / $7.00, about 30% below its official rate. A 50K input / 2K output analysis that costs $0.42 on Fable 5 costs just $0.21 on Opus 5 and $0.168 on Opus 5.5 — and because Opus 5.5's adaptive thinking is always on, allow for some output tokens to be used for thinking.

How to pay for the Claude API from Indonesia

Payments are processed by Stripe. There is no subscription and no need to set up billing in the Anthropic console: add funds to your balance, then each call deducts an amount based on the per-token rates above. Payment methods available from Indonesia:

• Credit / debit cards: Visa, Mastercard, American Express, JCB
• Apple Pay, Google Pay, Link (one-click payment)

Prices are set in USD. If the checkout displays a local currency, Stripe performs the conversion (the exchange rate includes a 2–4% conversion fee); if you pay in USD, your card issuer performs the conversion (check your bank's international transaction fees). Minimum top-up $10, prepaid, and the balance never expires. GoPay/OVO/DANA are not yet directly supported — but cards issued by Indonesian banks generally work; if declined, first enable international transactions in your banking app.

Cost examples

Real figures calculated using Kunavo rates:

Processing 10,000 support tickets per day with Haiku costs only about $12.60/day. Calculate it yourself:

# Tarif Claude di Kunavo (USD per 1M token): (input, output)
RATES = {
    "claude-haiku-4-5":  (0.70, 3.50),
    "claude-sonnet-4-6": (2.10, 10.50),
    "claude-opus-4-7":   (3.50, 17.50),
    "claude-opus-5":     (3.50, 17.50),
    "claude-opus-5-5":   (2.80, 14.00),
}

def biaya(model: str, in_tokens: int, out_tokens: int) -> float:
    i, o = RATES[model]
    return in_tokens / 1_000_000 * i + out_tokens / 1_000_000 * o

print(biaya("claude-haiku-4-5", 800, 200))      # satu tiket support   -> $0.00126
print(biaya("claude-sonnet-4-6", 6_000, 500))   # satu jawaban RAG     -> $0.0179
print(biaya("claude-opus-4-7", 50_000, 2_000))  # analisis konteks pjg -> $0.21
print(biaya("claude-opus-5", 50_000, 2_000))    # analisis Opus 5      -> $0.21
print(biaya("claude-opus-5-5", 50_000, 2_000))  # analisis Opus 5.5    -> $0.168

Indonesian uses slightly more tokens than English for the same content (~10–15%) — account for this when budgeting.

Prompt caching — the single biggest saving

If your system prompt or retrieval context is large and stable, prompt caching bills cached input at just 10% of the input rate (2.5% for Claude Fable 5.1 and 5% for Claude Opus 5.5). For long, repeated prompts, input costs drop by 60–90% — with just one additional field in the request. See the in-depth prompt caching guide and the Messages API reference.

One key for Claude, GPT, images, and video

One Kunavo key (sk-kn-…) calls Claude, GPT, and image, video, and audio models — all in an OpenAI-compatible format, charged against the same Stripe balance, with no need to reconcile invoices across vendors. Claude is available through both the compatible endpoint /v1/chat/completions and the native Anthropic API, Messages (/v1/messages), so your existing SDK keeps working. Find the steps to create a key in the authentication documentation.

Which Claude model should you choose?

Find code patterns for routing by difficulty in the AI cost optimization guide. Compare with OpenRouter: Kunavo vs. OpenRouter.

Claude in VS Code — Kilo Code, Cline, Roo Code

If you use a coding agent such as Kilo Code, Cline, or Roo Code, set the “OpenAI Compatible” provider to https://api.kunavo.com/v1 with your Kunavo key, then select a Claude model — just three fields. Agent sessions typically use thousands of tokens per step, so savings of 30% (depending on the model) add up quickly. See the full steps in the Kilo Code + Claude API guide.

Reducing Claude API costs

1. Start with prompt caching. Cached input is billed at 10% of the rate (2.5% for Claude Fable 5.1 and 5% for Claude Opus 5.5) — the biggest ROI.
2. Use a lower tier. Send easy tasks to Haiku; save Opus/Fable for the hardest parts.
3. Cap output. Set max_tokens — output costs about 5x more than input.
4. Trim input. Retrieve fewer, more relevant RAG chunks instead of the entire knowledge base.

Frequently asked questions