Claude Sonnet 5: specs, price, benchmarks and status

Claude Sonnet 5 specs at a glance

Here are the core numbers and platform details for claude-sonnet-5 from Anthropic's Sonnet 5 model page.

• Sonnet 5's 1M-token context window holds roughly 555k words on Claude's current tokenizer.
• You can run the model on the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry.
• On the Batch API, it can return up to 300K output tokens when you pass the output-300k-2026-03-24 beta header.

Claude Sonnet 5 API pricing

The table below breaks down the standard API, prompt caching and batch token rates for claude-sonnet-5.

When Sonnet 5 launched, Anthropic announced the $2/$10 rate as introductory pricing that would only last through August 31, 2026. Anthropic has since made $2/$10 the standard Sonnet 5 price, so the scheduled jump to $3/$15 per million tokens on September 1, 2026 will not happen.

If a workload sends 10 million input tokens and generates 2 million output tokens, you pay $20 for input plus $20 for output, coming out to $40 total (or $20 on the Batch API at 50% off). Check our Claude API pricing guide for other models' prices.

Sonnet 5 vs Sonnet 4.6 and Sonnet 5.5

Here is how claude-sonnet-5 stacks up against its predecessor claude-sonnet-4-6 and its successor Sonnet 5.5.

• Anthropic said Sonnet 5 performs close to Opus 4.8 and improves on Sonnet 4.6 across reasoning, tool use, coding and knowledge work.
• Sonnet 5.5 keeps the $2/$10 token price, cuts cache reads in half to $0.10 per million tokens and generates output 30%+ faster.
• In Anthropic's testing, Sonnet 5.5 delivered up to 30% lower cost per task than Sonnet 5.

Claude Sonnet 5 benchmarks

Anthropic published these claude-sonnet-5 benchmarks alongside Sonnet 5.5 and Opus 5.5 in its Sonnet 5.5 launch post on September 28, 2026.

Sonnet 5's weakest showing is in agentic terminal coding. It scores just 10.3% on Terminal-Bench 4.0, whereas Sonnet 5.5 jumps to 70.6% on the same test.

Anthropic also reports that running Sonnet 5.5 at Low or Medium effort beats Sonnet 5's best score on several benchmarks for roughly a tenth of the cost per task.

Which Claude plans include Sonnet 5

• At launch, Sonnet 5 was the default model on the Free and Pro plans, and available on Max, Team and Enterprise.
• Claude's pricing page lists Sonnet models as included on the Free, Pro and Max plans.
• Using Claude Code requires a Pro plan or above, as it is not included on Free.
• On the Claude API, developers call the model as claude-sonnet-5 and pay per token.

If you're still weighing model tiers, see our breakdown of Claude Sonnet vs Opus.

Should you still use Sonnet 5?

Sonnet 5's status is Active (legacy), meaning the model still runs on the API but receives no further updates. Anthropic recommends migrating to Claude Sonnet 5.5, though Sonnet 5 will not be retired any sooner than June 30, 2027, and Anthropic gives at least 60 days notice before retiring a model.

Because Sonnet 5.5 costs the same $2/$10 per million tokens, runs 30%+ faster and costs up to 30% less per task, it is the better default for new projects, while existing Sonnet 5 integrations can keep running without immediate changes. If your workload needs a heavier tier than Sonnet, look at Claude Opus.

Call Sonnet 5 from your Claude subscription

If you already pay for a Claude Pro or Max subscription, Plan2API lets you connect that account and call claude/claude-sonnet-5 through an OpenAI-compatible Chat Completions endpoint at https://api.plan2api.com/v1. Plan2API never bills you per token, though your Claude plan's own usage limits still apply.

Plan2API's Free plan is $0 (no card needed) and covers 1 connected subscription and 50 chat requests a day, while the Pro plan is $12 a month with no daily request cap. Read more about how the Claude subscription API works, or use the curl example that follows.

curl https://api.plan2api.com/v1/chat/completions \
  -H "Authorization: Bearer $PLAN2API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude/claude-sonnet-5",
    "messages": [
      {"role": "user", "content": "Summarize this changelog in 3 bullets."}
    ]
  }'