Better step-by-step answers
Does 'think step by step' help?
Ask the model to work through the problem step by step before it answers.
Why it matters
Multi-step problems are where models slip. Working through them first catches errors early.
How to use it
Add a line asking for the steps before the answer on any task with chained arithmetic or logic. Check that it helps on the model you use.
Holds
Pass rate up 20 points on Claude Haiku 4.5 and 34 points on GPT-5 mini. Inside the margin on Gemini 3.1 Flash Lite. Those results are tested via API. Also measured in Claude Code and Codex, reported separately on this page and never averaged with this.
tested via API · Holds on Claude Haiku 4.5, tested via API · Holds on GPT-5 mini, tested via API · No measured effect on Gemini 3.1 Flash Lite, tested via API
tested in Claude Code · No measured effect on Claude Fable 5, tested in Claude Code · No measured effect on Claude Opus 5, tested in Claude Code · Holds on Claude Haiku 4.5, tested in Claude Code · No measured effect on Claude Sonnet 5, tested in Claude Code · No measured effect on Claude Fable 5.1, tested in Claude Code
tested in Codex · No measured effect on GPT-5.4 mini, Codex run on GPT-5.4 mini, tested in Codex · No measured effect on GPT-5.6 Luna, Codex run on GPT-5.6 Luna, tested in Codex · No measured effect on GPT-5.6 Terra, Codex run on GPT-5.6 Terra, tested in Codex
What the marks mean
- Holds
- No measured effect
See the numbers per model
Without to with, per model
Via API tested via API
Claude Haiku 4.5
Holds
Gemini 3.1 Flash Lite
No measured effect
GPT-5 mini
Holds
deterministic pass rate, 0 to 1
In Claude Code tested in Claude Code
Claude Fable 5
No measured effect
Claude Fable 5.1
No measured effect
Claude Haiku 4.5
Holds
Claude Sonnet 5
No measured effect
Claude Opus 5
No measured effect
deterministic pass rate, 0 to 1
Show per-task detail
Every item, without and with, per model
Via API tested via API
- GPT-5 mini
- Claude Haiku 4.5
- Gemini 3.1 Flash Lite
deterministic pass rate, 0 to 1
Hover or focus an item to read its task and every model’s two values.
Pooled: the interval is taken over the records, not these means. These are those same records collapsed per item, so where an item carries more passes than another the lines need not average to the figure the table prints.
The item means behind this plot
| Item | GPT-5 mini | Claude Haiku 4.5 | Gemini 3.1 Flash Lite | |||
|---|---|---|---|---|---|---|
| Task | without | with | without | with | without | with |
| t01 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t02 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t04 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t05 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t07 | 0.000 | 1.000 | 0.000 | 1.000 | 0.000 | 1.000 |
| t09 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t03 | 0.200 | 1.000 | 0.000 | 1.000 | 1.000 | 1.000 |
| t06 | 0.200 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| t10 | 0.200 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| t08 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
In Claude Code tested in Claude Code
- Claude Haiku 4.5
- Claude Fable 5
- Claude Fable 5.1
- Claude Opus 5
- Claude Sonnet 5
deterministic pass rate, 0 to 1
Hover or focus an item to read its task and every model’s two values.
Pooled: the interval is taken over the records, not these means. These are those same records collapsed per item, so where an item carries more passes than another the lines need not average to the figure the table prints.
The item means behind this plot
| Item | Claude Haiku 4.5 | Claude Fable 5 | Claude Fable 5.1 | Claude Opus 5 | Claude Sonnet 5 | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Task | without | with | without | with | without | with | without | with | without | with |
| t01 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t02 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t04 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t05 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t07 | 0.000 | 1.000 | 0.000 | 1.000 | 0.000 | 1.000 | 0.000 | 0.600 | 0.000 | 1.000 |
| t09 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 | 0.000 |
| t03 | 0.200 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| t08 | 0.200 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| t06 | 0.600 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
| t10 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 | 1.000 |
Result
The phrase think step by step raised the pass rate by 20 points on Claude Haiku 4.5 and 34 points on GPT-5 mini tested via API while staying inside the margin on Gemini 3.1 Flash Lite, and showed no measured effect on Claude Fable 5, Claude Opus 5, Claude Sonnet 5 and Claude Fable 5.1; held on Claude Haiku 4.5 tested in Claude Code, and showed no measured effect on GPT-5.4 mini via the Codex run on GPT-5.4 mini, and showed no measured effect on GPT-5.6 Luna via the Codex run on GPT-5.6 Luna, and showed no measured effect on GPT-5.6 Terra via the Codex run on GPT-5.6 Terra. Measured 2026-08-12.
Show the per-model numbers
Claim tested: The phrase think step by step improves accuracy on multi-step word problems.
Holds on Claude Haiku 4.5 and GPT-5 mini. No measured effect on Gemini 3.1 Flash Lite.
Circulates in practitioner communities. Tested because it circulates, not because it is endorsed.
This tip is OpenAddict's plain-language read of the measured result. The measurement below is the evidence, and it is what the reading has to answer to.
Ledger idC05-think-step-by-step
Clarification, . Until 2026-08-31 this figure read 5, which was the pass count rather than a number of calls. What changed, and why it is filed once.
What was tested
This claim circulates in practitioner communities as advice about how to write prompts. That it circulates is an input to what gets tested here. It is a reason to test the claim, and it is not evidence for or against it. The result below is the evidence, and it is the only thing on this page that carries weight.
The comparison is paired. Two prompts differ in one respect, the manipulated variable, and are otherwise identical by construction. Nothing here supports a causal reading beyond that pairing.
Per-model numbers
| Measure | Claude Haiku 4.5 | GPT-5 mini | Gemini 3.1 Flash Lite |
|---|---|---|---|
| Control arm | 0.300n 50 | 0.160n 50 | 0.400n 50 |
| Treatment arm | 0.500n 50 | 0.500n 50 | 0.500n 50 |
| Delta | +0.200 | +0.340 | +0.100 |
| Interval, 95 percent | 0.010 to 0.390 | 0.166 to 0.514 | -0.096 to 0.296 |
| Orbit | Stable100 of 100 records | Stable100 of 100 records | In free drift100 of 100 records |
Orbit is assigned by the frozen status_v1 rule. On this scale, deterministic pass rate, 0 to 1, the pass threshold is +0.20 and the failure floor is -0.20, each requiring an interval that excludes zero.
In Claude Code
These cells are tested in Claude Code, on a subscription path with no API key. They are a second instrument and are never averaged with the figures above, which are tested via API. What that means, and how it was calibrated.
| Model | Control | Treatment | Delta | Interval | Pairs | Orbit |
|---|---|---|---|---|---|---|
| Claude Fable 5 | 0.4000 | 0.5000 | +0.100 | -0.096 to 0.296 | 50 | In free drift |
| Claude Opus 5 | 0.4000 | 0.4600 | +0.060 | -0.136 to 0.256 | 50 | In free drift |
| Claude Haiku 4.5 | 0.2000 | 0.5000 | +0.300 | 0.121 to 0.479 | 50 | Stable |
| Claude Sonnet 5 | 0.4000 | 0.5000 | +0.100 | -0.096 to 0.296 | 50 | In free drift |
| Claude Fable 5.1 | 0.4000 | 0.5000 | +0.100 | -0.096 to 0.296 | 50 | In free drift |
Method for this claim
- Task set
- 10 arithmetic word problems requiring two or three chained operations. Every answer was computed by hand and is a whole number. Problem text is identical between arms.
- Runs per model per arm
- 5 passes x 10 items
- Scoring
- Deterministic, via scoreExactMatch. A committed function scores each answer with no model in the loop.
- Pass criterion as written for the pilot
- Treatment exact-match rate exceeds control by at least 10 percentage points on the same 10 problems.
The published verdict comes from status_v1, not from the pass criterion above. The criterion is recorded because it is what the claim was registered with before the run.
Model versions, as recorded
| Model | Version string returned | Delta on this claim | Interval |
|---|---|---|---|
| Claude Haiku 4.5 | claude-haiku-4-5-20251001 | +0.200 | 0.010 to 0.390 |
| GPT-5 mini | gpt-5-mini-2025-08-07 | +0.340 | 0.166 to 0.514 |
| Gemini 3.1 Flash Lite | gemini-3.1-flash-lite | +0.100 | -0.096 to 0.296 |
Reading across models
Sampling was not held constant across vendors, so comparing one model column against another compares two settings as well as two models.
GPT-5 mini rejected the fixed sampling setting and ran at its own default on all 1,040 of its calls. The other models ran at temperature 0.