Claude Haiku 5.5 vs Claude Sonnet 5.5

Claude Haiku 5.5 costs a twentieth of Claude Sonnet 5.5 per token and ties it on FrontierCode 1.1, 46.4% against 46.2%. Sonnet 5.5 is far ahead on Terminal-Bench 4.0, 70.6% to 39.2%, so it stays the stronger agent in a terminal. Haiku's low price holds only for prompts up to 100K tokens.

First model Second model

Price: a twentieth, until the prompt passes 100K tokens

Haiku 5.5 costs $0.10 input, $0.50 output and $0.01 cached input per 1M tokens. A request whose prompt is over 100K tokens is billed at $0.50 and $2.50 instead. Sonnet 5.5 has one price: $2 and $10, $0.10 cached.

A coding agent resends the conversation with every call, so prompts grow through a session. Our feature session keeps them at 80K tokens and costs $0.54 on Haiku against $10.80 on Sonnet. A long agent day runs at 150K tokens, past Haiku's threshold, and costs $24.38 against $97.50: a quarter, not a twentieth.

What the scores say

Terminal-Bench 4.0, agentic coding in a real terminal: Sonnet 70.6%, Haiku 39.2%. FrontierCode 1.1, hard software engineering tasks: Haiku 46.4%, Sonnet 46.2% at max effort, a tie by any reading. Humanity's Last Exam with tools: Sonnet 64.5%, Haiku 57.4%.

Both read up to 1M tokens and write up to 128K. Anthropic released Sonnet 5.5 on 28 September 2026 and Haiku 5.5 nine days later, on 7 October.

Side by side

Claude Haiku 5.5 and Claude Sonnet 5.5: specifications and prices
	Claude Haiku 5.5	Claude Sonnet 5.5
Maker	Anthropic	Anthropic
Released	7 October 2026[1]	28 September 2026[2]
Context window	1M tokens[1]	1M tokens[2]
Max output	128K tokens[1]	128K tokens[2]
Input price	$0.10 per 1M tokens[1]	$2 per 1M tokens[2]
Output price	$0.50 per 1M tokens[1]	$10 per 1M tokens[2]
Cached input	$0.01 per 1M tokens[3]	$0.10 per 1M tokens[3]
Long prompts	Over 100K: $0.50 in, $2.50 out[1] Prompts over 100,000 tokens are billed at the higher rate.	One price
Price change	None announced	None announced
Open weights	No[1] Offered through the Claude API and cloud platforms; no weights are published.	No[2] Offered through the Claude API and cloud platforms; no weights are published.
Reads	text, image[1]	text, image[2]

Benchmarks

Scores as the makers published them, each with its source and date. Makers run these at different settings, so read a gap of a few points as a tie.

Published benchmark scores (higher is better)

Claude Haiku 5.5Claude Sonnet 5.5

Benchmarks: Claude Haiku 5.5 and Claude Sonnet 5.5
Benchmark	Claude Haiku 5.5	Claude Sonnet 5.5
Terminal-Bench 4.0Agentic coding tasks in a real terminal.	39.2%[4]	70.6%[5]
FrontierCode 1.1 (Main)Hard software engineering tasks.	46.4%[4]	46.2%[5] Max effort.
Humanity's Last Exam (with tools)Expert-level questions across many fields, tools allowed.	57.4%[4]	64.5%[5]

What a coding session costs

Our own estimate from the list prices above, for three sessions we define: a coding agent resends the conversation with every call, so prompts are large and grow. "Cached" assumes 90% of prompt tokens are read from the prompt cache and leaves out cache write fees. A long-prompt price applies when a call's prompt is over its threshold. Work out your own mix with the token cost calculator.

Feature session at list prices, US dollars (lower is cheaper)

Claude Haiku 5.5Claude Sonnet 5.5

Session costs: Claude Haiku 5.5 and Claude Sonnet 5.5
Session	Claude Haiku 5.5	Claude Haiku 5.5 cached	Claude Sonnet 5.5	Claude Sonnet 5.5 cached
Quick fix10 calls of 25K prompt and 1.5K output tokens	$0.03	$0.01	$0.65	$0.22
Feature session60 calls of 80K prompt and 2K output tokens	$0.54	$0.15	$10.80	$2.59
Long agent day300 calls of 150K prompt and 2.5K output tokens	$24.38	$6.15	$97.50	$20.55

Value for money and release gap

Claude Haiku 5.5 came out 9 days later. Price per point divides our feature-session cost by a published score: lower means more score for the money. It is a rough lens, not a ranking.

Value for money: Claude Haiku 5.5 and Claude Sonnet 5.5
Measure	Claude Haiku 5.5	Claude Sonnet 5.5
Input tokens per $1	10M	500K
Filling the whole context window once	$0.50	$2.00
Feature session per point of Terminal-Bench 4.0	$0.01	$0.15
Feature session per point of FrontierCode 1.1 (Main)	$0.01	$0.23
Feature session per point of Humanity's Last Exam (with tools)	$0.0094	$0.17

Questions

Is Claude Haiku 5.5 as good as Sonnet 5.5?

On FrontierCode 1.1 they tie, 46.4% against 46.2%. On Terminal-Bench 4.0 Sonnet 5.5 leads 70.6% to 39.2%, and on Humanity's Last Exam with tools 64.5% to 57.4%.

How much cheaper is Haiku 5.5 than Sonnet 5.5?

A twentieth per token for prompts up to 100K tokens: $0.10 against $2 per 1M input and $0.50 against $10 per 1M output. Above 100K, Haiku's rates are $0.50 and $2.50.

Do Haiku 5.5 and Sonnet 5.5 have the same context window?

Yes, 1M tokens each, and both write up to 128K tokens in one reply.

Most compared

Any other two models: pick them in the lists at the top of the page.

More tools

Sources

platform.claude.com: docs/en/models/haiku-5-5/overview, read 22 October 2026
platform.claude.com: docs/en/models/sonnet-5-5/overview, read 22 October 2026
platform.claude.com: docs/en/about-claude/pricing, read 22 October 2026
anthropic.com: claude-haiku-5-5, read 22 October 2026
anthropic.com: claude-sonnet-5-5, read 22 October 2026

Numbers from the makers' own pages, checked by hand, last read 22 October 2026. List prices in US dollars before tax. Spotted a mistake? office@shuriken.sh.

Made by Shuriken, the harness for coding agents. Its hot swapping moves your chats to another of your accounts or API keys when one nears its limit. Use only accounts you own and are entitled to use.

Try Shuriken on your own computer

Add your Claude, Codex and Gemini accounts once. Shuriken switches between them so your agents keep working.

Download Pricing

Coming soon