Requesty

Reasoning effort mix · September 2026

4.9% of chat requests set a reasoning effort in September 2026. Of those, "low" was the most requested level at 46.6%.

Share of requests that set a reasoning effort, and the mix of effort levels among those requests.

Data through September 30, 2026Updated November 11, 2026
This is the permanent snapshot for September 2026. See the latest month (October 2026)

Share of requests that set effort, September 2026

Source: Requesty production gateway
  • low46.6%
  • none17.5%
  • high14.5%
  • medium8.1%
  • token budget6.8%
  • max5.0%
  • minimal0.8%
  • xhigh0.7%
↓ JSON↓ CSV↓ Markdown↓ PNG

Key findings

  1. 01Low or no reasoning: 65.0% of requests that set an effort asked for none, minimal or low.
  2. 02High and above: 20.2% asked for high, xhigh or max.
  3. 03The share of requests that set an effort moved -3.9 pp since August 2026.

Share of requests that set effort by month

0.0%15.4%30.8%46.2%61.6%May 26Jun 26Jul 26Aug 26Sep 26low, May 26: 3.8%low, Jun 26: 4.2%low, Aug 26: 26.0%low, Sep 26: 46.6%none, May 26: 57.1%none, Jul 26: 8.0%none, Aug 26: 14.4%none, Sep 26: 17.5%high, May 26: 5.0%high, Jun 26: 3.8%high, Jul 26: 12.5%high, Aug 26: 45.2%high, Sep 26: 14.5%medium, May 26: 6.6%medium, Jun 26: 2.5%medium, Jul 26: 8.3%medium, Aug 26: 4.9%medium, Sep 26: 8.1%token budget, Jun 26: 5.3%token budget, Jul 26: 4.7%token budget, Aug 26: 1.5%token budget, Sep 26: 6.8%
lownonehighmediumtoken budget

What changed since August 2026

  • high45.2% → 14.5% (-30.7 pp)
  • low26.0% → 46.6% (+20.6 pp)
  • token budget1.5% → 6.8% (+5.3 pp)
  • medium4.9% → 8.1% (+3.2 pp)
  • none14.4% → 17.5% (+3.1 pp)

What to do about it

  • Set effort per task. Most traffic runs fine at low; keep high for planning and hard debugging.
  • Measure cost per task before and after changing effort. Higher effort raises output tokens.

Set reasoning effort once and Requesty maps it to each provider's format.

Read about reasoning

Data

EffortShare of requests that set effortOrganizations (band)
low46.60%100-499
none17.51%100-499
high14.47%100-499
medium8.08%100-499
token budget6.81%25-99
max4.97%25-99
minimal0.84%25-99
xhigh0.72%25-99

8 rows published. Rows below the floors are withheld.

Questions this answers

What reasoning_effort do developers use most?
4.9% of chat requests set a reasoning effort in September 2026. Of those, "low" was the most requested level at 46.6%.
How many API requests set a reasoning effort?
4.9% of chat requests set a reasoning effort in September 2026. Of those, "low" was the most requested level at 46.6%.
Do developers use high reasoning effort?
4.9% of chat requests set a reasoning effort in September 2026. Of those, "low" was the most requested level at 46.6%.

Method and definitions

  • Effort is read from the reasoning.effort or reasoning_effort request parameter. Numeric thinking budgets are grouped as token budget. Requests without the parameter count as not set.
  • A row is published only when it has at least 20 requests, at least 5 distinct organizations, and no single organization above 80% of its traffic. These are the same floors /rankings uses. Rows below the floors are withheld, not estimated.
  • Only shares, rates, medians and organization bands are published. Absolute request, token and spend totals stay private.
Set effort:
The request carried an explicit reasoning effort or thinking budget.

Caveats

  • Requests that do not set an effort run at the model default, which differs by model.
  • Some clients set an effort on every request even for models that ignore it.

Revision history and other months

← Aug 26Oct 26 →

Cite and reuse

Free to reuse under CC BY 4.0. Credit "Requesty Data" and link to this page. Cite the month URL so the numbers you quote stay fixed.

APA
Click to copy
BibTeX
Click to copy
CSV
Click to copy