ML/AIWork

For AI labs & DevRel

See where your model actually wins — in real hiring

Surveys ask developers what they like. API logs show your own traffic. Neither tells you which models employers are betting their roadmaps on. We measure that directly: the share of AI job postings naming each model, vendor and framework — by country, over time. Today OpenAI / GPT leads at 14.9%.

Open the public index →Request a vendor report

43,675 active AI postings · 18 countries · revealed preference

The model wars, demand side

Share of AI postings naming the model/vendorlive · 43,675 postings

OpenAI / GPT
14.9%
Claude (Anthropic)
13.3%
Gemini (Google)
6.4%
Llama
1.9%
Mistral
1.1%
BERT
0.5%

On the employer-interest side, Anthropic careers searches run at 1.9× OpenAI’s.

Mindshare is a black box — until you measure hiring

Surveys, API logs & press

  • Developer surveys measure sentiment, not commitment.
  • Your own API traffic is your own bubble — it can't see the field.
  • A funding headline isn't adoption.

Hiring demand

  • Employers pay for the stack they name in requirements.
  • Demand resolves by country and by quarter, not anecdote.
  • Your competitor's footprint stops being a guess.

The public index is free

The live scoreboard is published openly — cite it, embed it, link to it. It’s how labs, press and analysts find the number in the first place.

Open the public index →Employer Interest Index →

Why labs use it

  • See your real footprint vs the field — share of hiring demand, not your own logs
  • Track mindshare shifts by country and quarter — “+X pp this quarter”, not anecdotes
  • Prove DevRel impact with a number leadership trusts — tie programs to demand movement
  • Spot frameworks pulling demand toward — or away from — you — before it shows up in revenue
  • Own the narrative — the public index is free to cite and embed

How it works

  1. 1

    Watch the public index

    Live scoreboard of model/vendor demand — free, citable.

  2. 2

    Request a vendor report

    Your footprint, the frameworks and roles driving it, vs the field.

  3. 3

    Subscribe to comp-intel

    Quarterly tracking by country and segment for the trend line.

Pricing

The index itself stays free — that’s the point. What labs buy is the per-vendor read and the trend line over time.

Public index
Free

Live scoreboard. Cite it, embed it.

DevRel ROI report
$8K–15K

Your footprint, the drivers, the field — once.

Annual comp-intel
$15K–40K

Quarterly tracking by country & segment.

Request a vendor report →Open the public index

Questions

Is this stated or revealed preference?
Revealed. Job requirements are what employers commit budget and roadmap to — not what a developer says in a survey or which logo got press.
How is “naming a model” measured?
A dictionary match on word boundaries in postings (“Claude”, “GPT-4”, “Gemini”, “Llama”, “Mistral” and variants); a posting naming two vendors counts for both. Full rule on the methodology.
Can I track a framework, not just a model?
Yes — we track frameworks and tooling the same way (e.g. LangChain, vLLM), and can scope a report to the ones that move your demand.

Want your model's hiring footprint?

A per-vendor report shows where your model is named, who's driving it, and how you sit against the field.

Request a vendor report →See the free public index
n
43,675
as of
2026-06-14
version
v2026.06
geo
18 countries

Cite this

ML/AI Work, "ML/AI Work Model Demand Index", dataset v2026.06, n=43675, as of 2026-06-14. https://mlai.work/research/model-adoption

https://mlai.work/research/model-adoption