Does showing the work make AI better at big multiplication?

The question

Four-digit times four-digit numbers. Does asking for the working out first beat asking for just the answer?

What we found

• Haiku: 100% to 100% (no clear difference)
• Sonnet: 100% to 100% (no clear difference)

"Clear" means the two 95% ranges below do not overlap. When they overlap we say there is no clear difference, even if the percentages differ. With about 30 answers per cell, small gaps are noise.

Model	How we asked	Passed	Rate	95% range
Haiku	Asked for just the answer	30 of 30	100%	89% to 100%
Haiku	Told to work it out step by step	30 of 30	100%	89% to 100%
Sonnet	Asked for just the answer	30 of 30	100%	89% to 100%
Sonnet	Told to work it out step by step	30 of 30	100%	89% to 100%

What counts as a pass: the final number is exactly right.

What we saw reading the answers

• Every answer from both models was exactly right in both ways of asking, even for four-digit times four-digit numbers. We expected the straight answer to fail sometimes. It did not, so this test cannot tell the two ways apart.
• We do not know why. These models may use hidden reasoning before replying, or have seen these sizes of sum many times. Larger numbers or longer chains of arithmetic might show a gap.
• What to do: do not assume "think step by step" is needed for arithmetic on current Claude models. For money, still check with a calculator.

The exact wording

Shown with the first question as the example.

Asked for just the answer

System prompt: You are a helpful assistant.

What is 3472 x 5861? Reply with only the number.

Told to work it out step by step

System prompt: You are a helpful assistant.

What is 3472 x 5861? Work it out step by step, then finish with a line "Answer: N".

How we tested

• Models: Haiku (claude-haiku-5-5); Sonnet (claude-sonnet-5-5). Claude models only. We have not tested ChatGPT, Gemini, Grok or others, and results can differ.
• 10 different questions, each asked 3 times per way of asking, so the answer's randomness is included.
• Every answer was scored by a short program, not by a person's opinion. We then read a sample by hand.
• Each question was sent on its own through the Claude command-line tool, with tools, memory, settings and conversation history switched off, and no web access or examples. That tool still adds a short fixed note about the machine (operating system, working folder, the account's email domain), about 900 tokens in all. It is identical for both ways of asking, so comparisons stay fair, but absolute numbers may differ a little from a plain API call. We redacted the owner's company name from one published answer where a model repeated it.
• Run on 2026-10-26. Model versions change; treat results as a snapshot.

Limits

Ten multiplications of two four-digit numbers, no calculator tool. With a calculator tool both would be near perfect.

Raw data

Every question, answer and score is in the raw data file (filter by "exp": "multiply-steps"). Re-run it yourself with the harness in the project's testlab folder, or paste the prompts above into your own account.