Claude Sonnet 4.6 API

Claude Sonnet 4.6 is Anthropic's best balance of speed, intelligence, and cost — a versatile model for coding, agentic workflows, and everyday tasks with a 200K context window and 128K max output. Access it through EvoLink's unified API.

Using coding CLIs? Run Claude 4.6 Sonnet via EvoCode — One API for Code Agents & CLIs. (View Docs)

PRICING

PLAN	CONTEXT WINDOW	MAX OUTPUT	INPUT	OUTPUT	CACHE WRITE	CACHE READ
Claude Sonnet 4.6	200.0K	128.0K	≤200.0K$2.55-15% $3.00Official Price >200.0K$5.10-15% $6.00Official Price	≤200.0K$12.75-15% $15.00Official Price >200.0K$19.125-15% $22.50Official Price	≤200.0K$3.188-15% $3.75Official Price >200.0K$6.375-15% $7.50Official Price	≤200.0K$0.256-15% $0.300Official Price >200.0K$0.511-15% $0.600Official Price
Claude Sonnet 4.6 (Beta)	200.0K	128.0K	≤200.0K$0.780-74% $3.00Official Price >200.0K$1.56-74% $6.00Official Price	≤200.0K$3.90-74% $15.00Official Price >200.0K$5.85-74% $22.50Official Price	≤200.0K$0.975-74% $3.75Official Price >200.0K$1.95-74% $7.50Official Price	≤200.0K$0.078-74% $0.300Official Price >200.0K$0.156-74% $0.600Official Price

Web Search Tool

Server-side web search capability

$0.011/search

Pricing Note: Price unit: USD / 1M tokens

Cache Hit: Price applies to cached prompt tokens.

Two ways to run Claude Sonnet 4.6 — pick the tier that matches your workload.

· Claude Sonnet 4.6: the default tier for production reliability and predictable availability.
· Claude Sonnet 4.6 (Beta): a lower-cost tier with best-effort availability; retries recommended for retry-tolerant workloads.

Claude Sonnet 4.6 API — Anthropic's best-balanced model

Claude Sonnet 4.6 (Claude 4.6 Sonnet) delivers the ideal balance of intelligence, speed, and cost with a 200K context window and up to 128K output tokens for coding, agents, and complex workflows.

What can you build with the Claude Sonnet 4.6 API?

Versatile Coding Assistant

Use Sonnet 4.6 for day-to-day coding tasks — architecture, refactors, code review, and bug fixing. With up to 128K output tokens and a 200K context window, handle large codebases and generate comprehensive diffs, test suites, and implementation plans in a single request.

Try coding

Reliable Agent Workflows

Build agents that plan, call tools, and maintain context across multi-step tasks. Sonnet 4.6 balances intelligence and speed for agent-heavy workflows, delivering reliable tool use and consistent outputs at a fraction of flagship pricing.

Build agents

Extended Thinking & Analysis

Enable extended thinking for complex reasoning tasks. Sonnet 4.6 supports deeper analysis when needed while keeping costs predictable — ideal for research, planning, and technical strategy where you need more than a quick answer.

Explore reasoning

Why teams choose the Claude Sonnet 4.6 API on EvoLink

Get Anthropic's best-balanced model with stable model IDs, prompt caching, and unified routing through EvoLink's single API key.

Best balance of speed, intelligence, and cost

Sonnet 4.6 is purpose-built for teams that need strong performance across coding, analysis, and agentic tasks without flagship pricing.

128K max output for large-scale generation

Generate comprehensive code, documentation, and analysis in a single request — double the output capacity of previous models.

Cost control with prompt caching

Prompt caching supports 5-minute and 1-hour caches, and cache hits are billed at 0.1x the base input rate to reduce repeat costs.

How to integrate the Claude Sonnet 4.6 API

Connect through EvoLink, choose your model ID, and start building in minutes.

Step 1 — Create your EvoLink API key

Step 2 — Select the model ID

Use `claude-sonnet-4-6` to access the latest Sonnet 4.6 model through EvoLink's unified API.

Step 3 — Optimize quality and cost

Claude Sonnet 4.6 supports extended thinking for complex tasks and prompt caching to lower repeat costs — at $3/$15 per million tokens.

View API Docs

Claude Sonnet 4.6 API capabilities

Key specs and model features for production use

Context

200K Context Window

Read large documents or codebases in a single request without chunking.

Capacity

128K Max Output

Generate long-form answers, plans, and code without early truncation — double previous limits.

Intelligence

Extended Thinking

Enable deeper reasoning when tasks become complex, with predictable cost scaling.

Multimodal

Vision + Multilingual Input

Accept text and image inputs with strong multilingual understanding.

Efficiency

Prompt Caching Rates

Cache writes and reads are priced separately; cache hits are billed at 0.1x the base input price.

Reliability

Stable IDs & Aliases

Aliases auto-upgrade to the newest snapshot, while versioned IDs keep results consistent.

Claude Sonnet 4.6 API - FAQ

Everything you need to know about the product and billing.

Claude Sonnet 4.6 supports a 200K token context window and up to 128K output tokens in a single request, making it suitable for large codebases, long documents, and comprehensive generation tasks.

Use `claude-sonnet-4-6` to access the latest Sonnet 4.6 model. When a versioned snapshot becomes available, you can pin it for stable production behavior.

Base pricing is $3 per million input tokens and $15 per million output tokens (MTok = one million tokens). Prompt caching is billed separately with cache writes and cache hits priced per model. Provider pricing can differ on Bedrock or Vertex AI.

For Sonnet 4.6, cache writes are $3.75/MTok and cache hits are $0.30/MTok, which is 0.1x the base input price. Use caching for stable system prompts or repeated long context.

Sonnet 4.6 offers the best balance of speed, intelligence, and cost at $3/$15 per MTok, while Opus 4.6 is the flagship model at $5/$25 per MTok for the hardest tasks. Sonnet 4.6 also supports 128K max output vs Opus's 64K.

Claude Sonnet 4.6 is available via the Anthropic API and on AWS Bedrock and Google Vertex AI. EvoLink can route to the provider you choose.

Yes. All current Claude models support text and image input with multilingual capabilities, so you can combine documents, screenshots, and visuals in one request.

The models overview lists a reliable knowledge cutoff in May 2025 for Sonnet 4.6, with a broader training data cutoff in August 2025.