Back to catalog
Anthropic·TextHot30%+ off vs official

Claude Sonnet 4.6API

Balanced speed/quality — the everyday production workhorse, elite coding.

Input price

$2.10

per 1M tokens

Output price

$10.50

per 1M tokens

Single upstreamServed through one supplier today. If that supplier has an outage, calls fail until it recovers — failed calls are never billed. We're adding a second supplier to every text model.

Prompt caching

Pass cache_control on a stable prefix and cache hits bill at a fraction of input. Cache writes bill at the vendor's write rate — 1.25× input on Claude and GPT-5.6+, the input rate elsewhere.

Cache read
$0.21
Cache write
$2.625

Auto-derived from input × 0.1

Specs

Model ID
claude-sonnet-4-6
Endpoint
POST /v1/chat/completions
Category
Text
Provider
Anthropic
Context window (tokens)
1,000,000
Max output (tokens)
128,000
Capabilities
VisionTool callingStreamingReasoningLong context

Drop-in OpenAI compatibility

Point your existing OpenAI SDK at api.kunavo.com/v1 and swap the model id. No streaming changes, no SDK changes.

View docsClaude API pricing guide

Try it

Set KUNAVO_API_KEY from /app/keys, then run any of the snippets below.

curl
curl https://api.kunavo.com/v1/chat/completions \
  -H "Authorization: Bearer $KUNAVO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-4-6",
    "messages": [
      {"role": "user", "content": "Hello, Claude Sonnet 4.6"}
    ],
    "stream": false
  }'

Kunavo vs calling Anthropic directly

Same Claude Sonnet 4.6 weights, same responses. What a gateway changes is the account, the SDK and the bill — here is the honest side-by-side.

KunavoAnthropic direct
Price (per 1M tokens)$2.10 / $10.50 (−30%)$3.00 / $15.00
AccountOne Kunavo account, key in 2 minutes, $10 minimum top-upA Anthropic account plus its own billing setup
SDKKeep the OpenAI SDK — change base_url, set model to "claude-sonnet-4-6"Anthropic's SDK, or their OpenAI-compatible layer where offered
Same key also reachesClaude, GPT, image, video and audio modelsAnthropic's own catalog only
BillingOne Stripe balance, pay-as-you-go, never expires; failed requests unbilledA separate invoice per provider
Rate limitsShared gateway capacity, no contractual per-account limitAnthropic's own tier limits, raised by usage history or contract

Go direct to Anthropic when you need a contractual rate limit, an enterprise SLA, or a provider-only feature on day one. Use Kunavo when you want one key, one balance and a lower per-token rate across every provider.

Frequently asked questions

How much does Claude Sonnet 4.6 cost?

On Kunavo, Claude Sonnet 4.6 is $2.10 per 1M input tokens and $10.50 per 1M output tokens — about 30% under Anthropic's official price. Failed requests are never billed, and you pay as you go only for successful calls.

Can I call Claude Sonnet 4.6 with the OpenAI SDK?

Yes. Set base_url to https://api.kunavo.com/v1, use your Kunavo key, and set model to "claude-sonnet-4-6". Requests and responses are OpenAI-compatible; Claude is also available on the native /v1/messages endpoint

What endpoint does Claude Sonnet 4.6 use?

Call Claude Sonnet 4.6 through POST /v1/chat/completions on Kunavo’s OpenAI-compatible API.

What is Claude Sonnet 4.6 good for?

Claude Sonnet 4.6 is a text model from Anthropic, with support for Vision, Tool calling, Streaming, Reasoning, and Long context.

What is different from calling Anthropic directly?

The model is the same — about 30% under Anthropic's official price. The surrounding setup changes: one key and one Stripe balance replace a Anthropic account and its billing, you can use the OpenAI SDK by changing base_url, and the same key also reaches Claude, GPT, image, video and audio models. Gateway capacity is shared. If your use case requires provider-negotiated limits for Claude Sonnet 4.6, contact Anthropic directly.

Is Claude Sonnet 4.6 cheaper on Kunavo?

Yes. Claude Sonnet 4.6 is about 30% below Anthropic’s official list price on Kunavo, with no subscription or monthly minimum. You pay for successful usage, starting with a $10 top-up; failed requests are never billed and your balance never expires.

Related models