Skip to content
Open navigation
OpenOpen workspace

Stack Overflow Developer Survey 2026

LLM vendors: Used in agents by work setting

In the Remote group, 83% selected “Anthropic”; in the In-person group, 71% did. Of the 18 answers tested, “Anthropic” differs most by work setting, a negligible association: χ²(4, N = 6124) = 57.15, Holm-adjusted p < .001, V = .10. Respondents could select more than one answer, so the percentages add to more than 100%. These are unweighted respondents, not population estimates.

Published by crosstabs.com · Last updated ·
How results are checkedNamed statistical review pending. Selected automated reference cases are published on the validation page.

The table

The table

Question asked: Which LLM model vendors for AI tools have you used for work in the past year? Which would you like to use next year? Select all that apply. Respondents could select more than one answer. Shown by work setting, as the percentage of each group's respondents who selected each answer, with the number of respondents beneath. A respondent is counted in every row they selected, so a column adds to more than 100%.

LLM vendors: Used in agents by work setting: percentage of each group's respondents who selected each answer, and respondent counts
LLM vendors: Used in agentsHybrid (some in-person, leans heavy to flexibility) (A)Hybrid (some remote, leans heavy to in-person) (B)In-person (C)Remote (D)Your choice (very flexible, you can come in when you want or just as needed) (E)All
Anthropic82.3%C80780.4%C97970.9%53183.1%C1,92881.4%C69880.7%4,943
OpenAI55.5%54456.7%69155.5%41655.4%1,28555.4%47555.7%3,411
Google AI27.7%27127.3%33229.4%22027.1%62927.0%23227.5%1,684
Microsoft17.1%D16818.1%D22014.3%10712.7%29415.5%13315.1%922
DeepSeek8.6%848.4%10212.8%AB969.9%22910.1%879.8%598
Alibaba Qwen6.7%668.1%998.7%658.1%1889.6%828.2%500
Mistral AI4.7%464.2%514.4%334.4%1015.8%504.6%281
Moonshot AI3.5%343.9%473.2%244.2%984.4%383.9%241
Other, please specify2.7%262.9%353.5%263.9%915.4%AB463.7%224
xAI3.2%312.5%313.9%293.4%804.1%353.4%206
Zhipu AI1.9%192.5%303.1%232.2%503.5%302.5%152
Meta AI2.1%212.6%322.7%201.6%363.3%D282.2%137
Databricks (DBRX)1.2%121.3%160.8%60.9%202.0%d171.2%71
Snowflake0.7%70.7%90.9%71.0%232.1%b181.0%64
Cohere0.6%60.7%81.2%90.6%151.7%D150.9%53
Cascade0.4%40.6%70.5%40.5%121.3%110.6%38
Voyage AI (MongoDB)0.5%50.5%60.7%50.4%101.3%d110.6%37
Black Forest Labs0.5%50.5%60.5%40.3%81.4%D120.6%35
01.AI0.3%30.5%60.5%40.3%61.2%D100.5%29
Luel0.3%30.2%30.5%40.3%81.2%bd100.5%28
Stability AI0.2%20.2%30.7%50.2%51.4%ABD120.4%27
Guide Labs0.3%30.3%40.5%40.1%31.2%D100.4%24
Reka0.1%10.2%30.8%D60.1%31.2%ADb100.4%23
Respondents9801,2187492,3198586,124

A letter beside a percentage names the columns it is significantly higher than (uppercase 95%, lowercase 90%; pairwise column tests, Bonferroni-corrected). 24,779 of 30,903 respondents have no answer to one of the two questions (not asked or not answered) and are left out.

Analysis

Results

Chi-square test of “Anthropic”

χ² = 57.15

df = 4, Holm-adjusted p < .001

Effect size

Cramér's V = 0.097

a negligible association

The strongest answer, APA-style: χ²(4, N = 6124) = 57.15, Holm-adjusted p < .001, V = .10. Each answer is tested on its own, as the respondents who selected it against those who did not, across the 5 groups (N = 6,124). The p-values are Holm-adjusted for the 18 answers tested; 2 of them differ by work setting at the 0.05 level. 5 answers are not tested: too few respondents selected them for the test to be reliable (an expected count below 1, or more than 20% of the expected counts below 5).

LLM vendors: Used in agents by work setting: the test of each answer, selected against not selected, by work setting
AnswerSelectedχ² (df = 4)Adjusted pCramér's V
Anthropic80.7%57.15< .0010.097
OpenAI55.7%0.671.0000.010
Google AI27.5%1.621.0000.016
Microsoft15.1%22.68.0020.061
DeepSeek9.8%12.34.2110.045
Alibaba Qwen8.2%5.171.0000.029
Mistral AI4.6%3.831.0000.025
Moonshot AI3.9%2.711.0000.021
Other, please specify3.7%12.54.2110.045
xAI3.4%4.631.0000.027
Zhipu AI2.5%6.94.9750.034
Meta AI2.2%10.63.3410.042
Databricks (DBRX)1.2%7.97.8560.036
Snowflake1.0%11.49.2810.043
Cohere0.9%11.42.2810.043
Cascade0.6%7.34.9530.035
Voyage AI (MongoDB)0.6%8.17.8560.037
Black Forest Labs0.6%12.64.2110.045
01.AI0.5%Not tested: expected counts too small
Luel0.5%Not tested: expected counts too small
Stability AI0.4%Not tested: expected counts too small
Guide Labs0.4%Not tested: expected counts too small
Reka0.4%Not tested: expected counts too small

How to read it

No single chi-square test covers this table. A respondent appears in every row they selected, so the rows are not separate groups of people, and a test of the whole table would count the same person several times. Each answer is tested on its own instead: did the share who selected it differ between the groups? Testing 18 answers gives 18 chances of a false alarm, so the p-values are Holm-adjusted to keep the chance of any false alarm in this table at 0.05.

For “Anthropic” the test rejects independence at the 0.05 level (Holm-adjusted p < .001): differences this large between the groups are unlikely if selecting it did not depend on work setting. Cramér's V of 0.097 is in the negligible range, which is trivially weak in practical terms, even where a sample this large can detect it.

With 6,124 respondents, even small differences reach statistical significance. Read the percentages and the effect size before the p-value.

Limits

  • The survey has no weights. These tables are unweighted counts of the people who answered, so they describe the respondents, not developers in general.
  • The test treats respondents as a simple random sample. They are people who chose to answer a survey Stack Overflow promoted, not a random sample of developers, so a result holds for these respondents and may not hold beyond them.
  • The adjustment covers the answers of this question by this group only, not the other tables in the library. The headline answer is the strongest of those tested, so its effect size is larger than a typical answer's.
  • A difference between groups is not evidence that group membership causes the answer.

Next steps

Work with this table

Open the test of “Anthropic” in the calculator

The link fills in the respondents who selected this answer and those who did not, by group. The whole table does not open there: its rows overlap, and the calculator's test needs rows that are separate people.

Open this question in the full file

The workspace opens all 30,903 respondents in your browser with this question selected. The survey has no weights, so the table there is unweighted too. Change the group, add a filter or export from there.

Open in the workspace →

Run this on your own data — free, no signup

Upload a CSV, Excel, SPSS, Stata, SAS, R, Parquet or JSON file. Everything runs in your browser; your file never leaves your device.

Open the workspace →

The same question by other groups

Other questions that differ by work setting

Source

Stack Overflow. 2026 Developer Survey, https://survey.stackoverflow.co/2026, licensed under the ODbL 1.0. Tabulated by crosstabs.com. Variables LLMUsedInAgents (LLM vendors: Used in agents) and WorkSetting. Stack Overflow licenses the survey data under the Open Database License (ODbL) 1.0 and its contents under the Database Contents License (DbCL) 1.0: credit Stack Overflow, keep the licence notice, and share any database made from the data under the ODbL. These tables and their downloads are offered on those terms.

  1. [1]Stack Overflow: 2026 Developer Survey public results (survey_results_public.csv)survey.stackoverflow.co
  2. [2]Stack Overflow: Open Database License (ODbL) 1.0opendatacommons.org