# Anthropic launches Claude Sonnet 5.5: benchmarks, pricing and updates

> Claude Sonnet 5.5 vs. Sonnet 5 and Opus 5.5: benchmark charts, API pricing, and tasks we propose testing with marketing teams.


Published: 2026-09-28. Original source: [Anthropic · Introducing Claude Sonnet 5.5](https://www.anthropic.com/claude-sonnet-5-5).

**Anthropic released Claude Sonnet 5.5 on September 28.** In the company's tests, it generates text more than 30% faster and can reduce cost per task by up to 30% compared with Sonnet 5.

The cover image is from Anthropic. See the [original announcement](https://www.anthropic.com/claude-sonnet-5-5) and [Anthropic's post on X](https://x.com/AnthropicAI/status/2104633259925630995).

## Sonnet 5.5 benchmarks vs. Sonnet 5 and Opus 5.5

Select an evaluation to compare the published results. Each chart retains its scale and unit: **Elo for professional work; percentages for the other tests**.

For a marketing team, each test answers a different question: Chartography measures chart interpretation; CursorBench evaluates code changes. We can use them to choose which tasks to try with the model. Campaign copy quality needs its own evaluation.

## How much does Claude Sonnet 5.5 cost?

[Official API pricing](https://platform.claude.com/docs/en/about-claude/pricing) is **US$2 per million input tokens**, **US$10 per million output tokens**, and **US$0.20 per million cache-read tokens**. Per-token prices are unchanged from Sonnet 5. The reported savings come from using fewer tokens to finish the work. A Claude subscription is billed separately.

At those rates, a call with 10,000 input tokens and 2,000 output tokens costs **US$0.04**, excluding caching and tool charges. This is a calculated example: a report requiring several calls will have different usage.

## How to access the Sonnet 5.5 API

The model ID is `claude-sonnet-5-5`. The [model documentation](https://platform.claude.com/docs/en/models/overview) lists access through the Claude Platform, AWS, Google Cloud and Microsoft Foundry, along with a one-million-token context window. For an existing integration, consult the [migration guide](https://platform.claude.com/docs/en/models/sonnet-5-5/migration-guide) before switching models.

## What to try with Sonnet 5.5 in marketing

We suggest starting with deliverables the team can assess using its own materials:

1. **Analyze posts:** compare interactions on carousels from the same account and period, with the values used.
2. **Prepare a draft:** supply a brief and request copy for each slide, checking whether it preserves product facts.
3. **Fix a landing page:** specify a concrete problem and verify the change in a mobile preview before publishing.

With [HeyMark connected to Claude](/en/guides/connect-claude-heymark/), you can ask the assistant for your accounts' posts and metrics to prepare that comparison. **We have not yet tested Sonnet 5.5 with HeyMark**; these are the use cases we propose evaluating.

Our guide to [analyzing social media with real data](/en/guides/analyze-social-media-chatgpt-real-data/) shows how to request the figures and check the conclusions.

## Plan your next post in HeyMark.

Keep the idea, review the draft with your team, and see how it performed in the accounts you connected.

[Start free](https://app.heymark.ai) · [See how it works](https://heymark.ai/en/#product)

## Developer and agent resources

- [HeyMark MCP documentation](https://heymark.ai/en/mcp/)
- [llms.txt](https://heymark.ai/llms.txt)
- [Full site content for language models](https://heymark.ai/llms-full.txt)
- MCP protocol endpoint: `POST https://mcp.heymark.ai`
- OAuth protected-resource metadata: [/.well-known/oauth-protected-resource](https://mcp.heymark.ai/.well-known/oauth-protected-resource)
- MCP server card: [/server-card](https://mcp.heymark.ai/server-card)