Skip to content
Escape
Meridian

The future is still open.

· Oct 5, 2026

Epoch AI charts rising coding-agent spending among OpenAI researchers

Markdown · JSON · Original source

’s page is dated October 5, 2026. Epoch AI describes extracting data from two charts in ’s post “Research acceleration: The view inside OpenAI,” covering researchers’ coding-agent use from January to mid-August 2026. Epoch AI fit trends for median and 90th percentile researchers, preferring an exponential trend with one breakpoint in the slope. It reports recent monthly growth rates of 1.8× and 2.2×, respectively, equivalent to doubling times of 34 and 27 days. Both series, Epoch AI says, grew several thousandfold from their first positive January values to mid-August. The analysis concerns spending measured in USD at prices, not a direct demonstration of research outcomes or productivity. Its significance is the reported pace of spending growth among OpenAI researchers; the evidence here does not establish what that spending produced or whether the trend extends beyond them.

Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.

Epoch AIArticle · epoch.aiOpen original source

Original source

· Article

Source details

Original source date: Oct 5, 2026 (2026-10-05T00:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-05T05:33:23.869Z. First observed: 2026-10-18T02:42:00.209Z. Retrieved: 2026-10-23T22:53:40.153Z.

Referenced by

Recorded citations and public file attachments. A reference alone does not establish agreement or influence.

No public references recorded here yet.

Reference data (JSON)

Limits

Text as retrieved on October 23, 2026; originally published October 5, 2026. says its analysis extracts data embedded in ’s first two charts; the retrieved text does not include the underlying data or the details of the assumptions and limitations.

Open question

How do the spending trends extracted from ’s charts compare with the underlying usage data?

Source-stated

Epoch AI says it extracted data from OpenAI’s first two charts about researchers’ coding-agent use from January to mid-August 2026.

We extracted the data embedded in OpenAI’s first two charts and fit trends to show how quickly usage grew for median and 90th percentile researchers.

Source-stated

Epoch AI’s preferred fit is an exponential trend with one breakpoint in the slope.

Our preferred fit is an exponential trend with a single breakpoint in the slope.

Source-stated

Epoch AI reports monthly growth rates of 1.8× and 2.2× for median and 90th percentile researchers, respectively, and doubling times of 34 and 27 days.

In recent months, median and 90th percentile researchers have seen similar growth rates of 1.8× and 2.2× per month, or doubling times of 34 and 27 days, respectively.

Source-stated

Epoch AI says both series grew several thousandfold between their first positive January values and mid-August.

Both series grew several thousandfold from their first positive January values to mid-August.

Source and contributor context

Epoch AI · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

API

An application programming interface: a defined way for software to request data or actions from another system.

Epoch AI

Epoch AI studies trends in AI capability, training resources and the infrastructure behind them.

OpenAI

OpenAI develops AI models and products and publishes research, release announcements and evaluations.

Editorial provenance

Version 1. Publication ID: 427cdea7-ba9d-ee94-0ec7-40c70252250e.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T05:28:47.253Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: 3cb434bf6e181239c1db689c9b45cfa643f40b052bcd8495dd5fc2c7e54ec478. Source text hash: c6d024130b74fb29cd46f89807da263af2ce8a704f5e0e5fad736afa2c4b1b94.

Revision and source-check history

Version 1 · 2026-11-05T05:33:23.869Z

Epoch AI charts rising coding-agent spending among OpenAI researchers

’s page is dated October 5, 2026. Epoch AI describes extracting data from two charts in ’s post “Research acceleration: The view inside OpenAI,” covering researchers’ coding-agent use from January to mid-August 2026. Epoch AI fit trends for median and 90th percentile researchers, preferring an exponential trend with one breakpoint in the slope. It reports recent monthly growth rates of 1.8× and 2.2×, respectively, equivalent to doubling times of 34 and 27 days. Both series, Epoch AI says, grew several thousandfold from their first positive January values to mid-August. The analysis concerns spending measured in USD at prices, not a direct demonstration of research outcomes or productivity. Its significance is the reported pace of spending growth among OpenAI researchers; the evidence here does not establish what that spending produced or whether the trend extends beyond them.

Original source

· Article

Source details

Original source date: Oct 5, 2026 (2026-10-05T00:00:00.000Z).

Date basis: Date supplied by the publisher’s feed.

Added to Escape Meridian: 2026-11-05T05:33:23.869Z. First observed: 2026-10-18T02:42:00.209Z. Retrieved: 2026-10-23T22:53:40.153Z.

Limits

Text as retrieved on October 23, 2026; originally published October 5, 2026. says its analysis extracts data embedded in ’s first two charts; the retrieved text does not include the underlying data or the details of the assumptions and limitations.

Open question

How do the spending trends extracted from ’s charts compare with the underlying usage data?

Source-stated

Epoch AI says it extracted data from OpenAI’s first two charts about researchers’ coding-agent use from January to mid-August 2026.

We extracted the data embedded in OpenAI’s first two charts and fit trends to show how quickly usage grew for median and 90th percentile researchers.

Source-stated

Epoch AI’s preferred fit is an exponential trend with one breakpoint in the slope.

Our preferred fit is an exponential trend with a single breakpoint in the slope.

Source-stated

Epoch AI reports monthly growth rates of 1.8× and 2.2× for median and 90th percentile researchers, respectively, and doubling times of 34 and 27 days.

In recent months, median and 90th percentile researchers have seen similar growth rates of 1.8× and 2.2× per month, or doubling times of 34 and 27 days, respectively.

Source-stated

Epoch AI says both series grew several thousandfold between their first positive January values and mid-August.

Both series grew several thousandfold from their first positive January values to mid-August.

Source and contributor context

Epoch AI · Publisher

The publisher is identified here. No individual author has been inferred.

Source media

Media retrieval status: unchecked.

Terms and references

Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.

API

An application programming interface: a defined way for software to request data or actions from another system.

Epoch AI

Epoch AI studies trends in AI capability, training resources and the infrastructure behind them.

OpenAI

OpenAI develops AI models and products and publishes research, release announcements and evaluations.

Editorial provenance

Version 1. Publication ID: 427cdea7-ba9d-ee94-0ec7-40c70252250e.

Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.

Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-05T05:28:47.253Z. This checks fidelity to the source, not whether its claims are independently true.

Editorial selection: accept; reasons: bridge_admitted_by_model.

Observation: 3cb434bf6e181239c1db689c9b45cfa643f40b052bcd8495dd5fc2c7e54ec478. Source text hash: c6d024130b74fb29cd46f89807da263af2ce8a704f5e0e5fad736afa2c4b1b94.

uncertain · 2026-11-05T05:28:47.253Z · anthropic/claude-opus-5.5

Publication: 427cdea7-ba9d-ee94-0ec7-40c70252250e

attribution — pass: This is a single Epoch AI data-insight page. The reading attributes the extraction, fitting and growth figures to Epoch AI, and attributes the underlying charts to OpenAI's post.

We extracted the data embedded in OpenAI’s first two charts and fit trends to show how quickly usage grew for median and 90th percentile researchers.

summary — uncertain: The summary is faithful and appropriately cautious. However, the central term "coding agent" is never explained, and it is not in the definitions list (only API is). A non-specialist may not know these are AI tools that write and execute code for users. Smallest repair: add a term-library entry for "coding agent" (and possibly "breakpoint"). The question is also somewhat unclear, since the extracted data is the chart data; it could ask what the spending growth implies or leaves unknown.

date — uncertain: The reading presents October 5, 2026 as the page date, matching originalPublished. That value is a feed backfill, though, and the page's downloadable CSV is marked updated Sep. 15, 2026, which suggests the insight may have first appeared earlier. Nothing post-publication or hindsight is used, and this is a publisher page, so no archive disclosure is needed.

CSV, Updated Sep. 15, 2026

support — pass: Each claim matches the source text verbatim in substance: the extraction from OpenAI's first two charts, the preferred breakpoint exponential fit, the 1.8× and 2.2× monthly rates with 34- and 27-day doubling times, and the several-thousandfold growth. The USD-at-API-prices framing comes from the data label. The caveats about outcomes and generalization are appropriately limited.

Weekly mean daily coding-agent spending per OpenAI researcher, median and 90th percentile (USD at API prices)

media — pass: No media was selected and no verified candidates were supplied, so there is no media identity or relevance issue.

Challenge or correct this reading

Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.