Frontier training compute growth (doubling time)
1What this measures
Doubling time of frontier models' training compute, fitted log-linearly against publication date from 2020 on Epoch's frontier-model table. Derived metric training_compute_doubling_days.
Why it matters. The invention stage's own speed limit. The brief's reading is 4-5x a year; a slowdown here is the first thing the data-and-compute constraint would show.
- Proxy types
- benchmark, model_release
- Unit
- days
- Cadence
- annual
- Valve
- invention to product
2How we track this
- derived metric
training_compute_doubling_days(formula in the semantic layer)
- Normal band
- ≥ 300 days
- Fast band
- ≤ 182 days
- Falsifying
- —
Normal = doubling no faster than every 300 days (about 2.3x a year, the pre-2010 pace of compute growth); fast = every 182 days or less (4x a year or more, Epoch's 2010-2025 trend). Between is emerging.
Applied to metric:training_compute_doubling_days.
3Tracker interpretation
Fitted on Epoch's estimates, most of which are inferred for closed models; the newest models are missing until Epoch estimates them.
4Evidence
66 observations. Hollow points are disputed (see counterevidence). Every point links to its observation.
Derived rows (1)
| as of | dims | value | inputs |
|---|---|---|---|
| 2025-07-09 | -10.0 38 0.884 | 170 days | obs:01535b14obs:032a5a5dobs:04b7374f+63 more |
5Status and reasoning
Log-linear fit on Epoch's 38 frontier models published since 2020: training compute doubles every 170 days (95% CI 152-193), inside the fast band (182 days or less, 4x a year or more) and matching Epoch's own 4-5x reading. Single tier-6 source; the estimates are inferred for most closed models, so the status is capped at emerging. Latest frontier model with an estimate: July 2025. Initial seed.
The tracker's prior expectation was faster than normal; the evaluator reads emerging. The evaluator wins until a reviewed override.
6Timeline notes
- 2025-07-09 · 170 daysas of 2025-07-09(66 obs)
7Counterevidence
What cuts against this reading
Compute estimates for closed models carry wide uncertainty; the frontier set is Epoch's selection; training compute is an input, not a capability.
8Update history
- 2026-09-10unmeasured to emergingconf — → 60 · evaluate
Log-linear fit on Epoch's 38 frontier models published since 2020: training compute doubles every 170 days (95% CI 152-193), inside the fast band (182 days or less, 4x a year or more) and matching Epoch's own 4-5x reading. Single tier-6 source; the estimates are inferred for most closed models, so the status is capped at emerging. Latest frontier model with an estimate: July 2025. Initial seed.
9Confidence
60 / 95 — mixed or hard to operationalise
Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; below 50 limited or vague.
10Related
- Training power draw growth (doubling time) emerging
- METR 50% time horizon faster than normal
- Crosswalk: Methods ⇄ Model (same valve)