Claude Sonnet 5.5 pricing, $2 in and $10 out per 1M tokens

Anthropic says Claude Sonnet 5.5 offers the best combination of speed and intelligence, built for everyday coding, agent and enterprise workloads. It costs $2 per 1M input tokens and $10 per 1M output tokens, with a 1,000,000-token context window and 128,000-token max output. This page gives the official price grid, the Batch API rates at half price, the cache prices (cache hits are only 5% of the base input price on Sonnet 5.5), and a calculator for estimating your workload cost.

Snapshot date is 8 November 2026. Each Anthropic list price on this page, in USD per 1M tokens, is from the Claude pricing page, fetched that day. Limits, the release date and the model ID are from the Claude Sonnet 5.5 overview. The worked examples are KickLLM calculations from those list prices. Prices change, so treat this page as a dated snapshot.

Claude Sonnet 5.5, model ID claude-sonnet-5-5 on the Claude API, was released on September 28, 2026. Its price is flat across the entire 1M-token context window. Unlike Claude Haiku 5.5, it has no prompt-length step, so a 900,000-token request is billed at the same per-token rate as a 9,000-token request. Prompt caching and the Batch API discount apply at their standard rates across the full context. The default effort level is high with adaptive thinking. See Claude Sonnet 5.5 overview and Claude pricing, long context.

$2input per 1M tokens
$10output per 1M tokens
1,000,000token context window
128,000token max output

The full official price grid

USD per 1M tokens. See Claude pricing, model pricing, snapshot 8 November 2026

Price lineClaude Sonnet 5.5How it relates to base inputSource
Base input$21xModel pricing
Cache write, 5-minute$2.501.25xModel pricing
Cache write, 1-hour$42xModel pricing
Cache hit (read)$0.100.05x, half the usual 0.1x ratePrompt caching
Output$105xModel pricing
Batch API input$10.5xBatch processing
Batch API output$52.5xBatch processing
US-only inference, inference_geo set to "us"Input, output and cache lines × 1.11.1x of each of those linesData residency pricing
Prompts up to the 1M token context windowSame rates, no long-prompt step1xLong context pricing

On most Claude models a cache hit costs 0.1x the base input price. On Claude Opus 5.5 and Claude Sonnet 5.5 it costs only 0.05x, $0.10 per 1M tokens against the $2 base. After the cache write, repeated reads of a large system prompt or codebase context cost one twentieth of the uncached input price. See Claude pricing, prompt caching, snapshot 8 November 2026. The minimum cacheable prompt is 512 tokens. See Claude Sonnet 5.5 overview, snapshot 8 November 2026.

Batch API prices

Batch API rates, USD per 1M tokens. See Claude pricing, batch processing, snapshot 8 November 2026

Batch price lineClaude Sonnet 5.5Claude Sonnet 4.6
Batch input$1$1.50
Batch output$5$7.50

The Batch API gives a flat 50% discount on both input and output tokens, in exchange for asynchronous processing. That is $1 per 1M input and $5 per 1M output on Sonnet 5.5, against $1.50 / $7.50 on the older Claude Sonnet 4.6. Prompt caching multipliers stack with the batch discount, and the batch rates apply across the full 1M context window with no long-prompt step. See Claude pricing, batch processing, snapshot 8 November 2026.

US-only inference. The Claude pricing page says that for Claude 4.6 and later models, setting inference_geo to "us" multiplies every token price (input, output, cache writes and cache reads) by 1.1. This applies on the Claude API, Claude Platform on AWS and Claude Models on Google Cloud (Anthropic managed), and to US Data Zone Standard deployments in Microsoft Foundry. Amazon Bedrock and Claude Models on Google Cloud (Google managed) have their own regional pricing. See Claude pricing, data residency, snapshot 8 November 2026.

Worked cost examples

KickLLM calculations at the official snapshot rates

A 30 day workload with no cache

The job uses 3,000,000 input tokens and 600,000 output tokens a day, which is 90,000,000 input and 18,000,000 output tokens a month. At standard rates the bill is $180 of input plus $180 of output, $360 in total. On the Batch API the same job costs $90 plus $90, $180 in total. The bills are the same on Claude Sonnet 5 because this job has no cache reads.

The same job with a 1 hour cache

Now 2,000,000 of the daily input tokens go through a 1 hour cache at a 95 percent hit rate. Each month has 30,000,000 uncached input tokens, 3,000,000 cache writes, 57,000,000 cache hits and 18,000,000 output tokens. Claude Sonnet 5.5 bills $60, $12, $5.70 and $180, which is $257.70, or 28.42 percent below the uncached $360. Claude Sonnet 5 bills $11.40 for the same hits and $263.40 in total. For this mix the difference is $5.70 a month, 2.16 percent of the Sonnet 5 bill.

One long context request

A request with 900,000 input tokens and 4,000 output tokens costs $1.80 of input plus $0.04 of output, $1.84 in total, with no long context surcharge.

When a cache write pays back

A 5 minute write costs 1.25x of base input and pays back after 1 read, because 1.25x plus 0.05x is 1.30x against 2x uncached. A 1 hour write costs 2x and pays back after 2 reads, because 2x plus 0.10x is 2.10x against 3x uncached. After 1 read it is still 2.05x against 2x.

Sonnet 5.5 cost calculator

KickLLM estimate in USD per request and per workload, at the standard or Batch API rates that you select. See Claude pricing for prices and Sonnet 5.5 overview for limits. Snapshot 8 November 2026

Check with the token counts of the Anthropic Managed Agents worked example (Claude pricing, fetched 8 November 2026). That example uses Claude Opus 5.5, with a token part of $0.50 for 50,000 input and 15,000 output tokens, and $0.348 with 40,000 of the input as cache reads. The Claude Sonnet 5.5 input, output and cache hit rates are half of the Opus 5.5 rates, so the same counts give $0.25 and $0.174 here. To see these totals, enter 50,000 uncached input and 15,000 output, then 10,000 uncached input, 40,000 cache read and 15,000 output.

KickLLM estimate, not an Anthropic figure. The documented limits are the 1,000,000-token context window (prompt + output) and 128,000 max output tokens for Sonnet 5.5. The Batch option uses the synchronous 128,000 output limit; the 300,000-token Batch output with the output-300k-2026-03-24 beta header is not modeled. Batch cache prices are derived (cache multiplier × 50% batch discount); the official batch table lists only input and output. US-only with Batch is a KickLLM derivation (1.1x applied to the batch rates). See Claude pricing, Claude Sonnet 5.5 overview, Context windows (the window counts the prompt and the output), fetched 8 November 2026.

Claude Sonnet 5.5 against Claude Sonnet 5

Price lineClaude Sonnet 5.5Claude Sonnet 5
Base input$2$2
Cache write, 5 minutes$2.50$2.50
Cache write, 1 hour$4$4
Cache hit (read)$0.10$0.20
Output$10$10
Batch API input$1$1
Batch API output$5$5

The what is new page says that Claude Sonnet 5.5 has the same prices as Claude Sonnet 5, except for prompt cache reads. The same page says that the tokenizer is the same as that of Claude Sonnet 5, so the same text produces the same token counts. A workload with no cache reads therefore has the same token bill on both models. Any difference between the two bills comes from the cache read tokens.

Where these numbers come from

Calculator formula, per request. Cost = (uncached input × input rate + cache reads × cache read rate + 5-minute writes × 5-minute write rate + 1-hour writes × 1-hour write rate + output × output rate) / 1,000,000. Standard rates are $2, $0.10, $2.50, $4 and $10. Batch rates are $1 and $5 for input and output, and half of each cache rate. US-only multiplies the total by 1.1. Workload total = cost per request × requests. The Sonnet 5 line uses the same formula with the Sonnet 5 standard rates $2, $0.20, $2.50, $4 and $10. The Batch line uses the same token counts at the Batch rates, and it shows only when Standard API is selected.

The main sources, all fetched as text on 8 November 2026, are the Claude pricing page, which gives the base price grid, the cache write and hit lines (including the 0.05x Sonnet 5.5 cache-hit rate), the Batch prices, the data residency multiplier and the long-context rule; and the model pages, the Claude Sonnet 5.5 overview, which gives the context window, max output, release and retirement dates and the platform model IDs, and What's new in Claude Sonnet 5.5, which lists the breaking changes. The Claude Sonnet 5 rates in the comparison table are from the same Claude pricing page.

The calculator and worked examples are KickLLM estimates. They are not Anthropic figures, and your invoice can be different. KickLLM is not affiliated with Anthropic or any other model provider.

KickLLM Margin Studio · Offline analysis app

Know your AI costs. Now plan your margin.

Turn your usage CSV into a cost breakdown, test revenue and growth assumptions, and export a report for your next pricing decision. Margin Studio runs locally with rates you supply.

Get Margin Studio, $39 Try the interactive preview → One-time purchase · Downloadable ZIP
By the same builder GitHub · theluckystrike BeLikeNative · Grammar AI EarlyThunder · Dev Blog Bug Bounty Reality Zovo · AI Dev Tools