Last updated: October 2, 2026
This is a historical review of 2025, not a forecast for a year that is still ahead. If you are assessing an old gold price prediction, first match its date, currency, benchmark and forecast horizon to the result being measured.
Content reviewed: October 2, 2026. Historical data below cover 2025.
What actually happened in 2025?
| Measure | Reported result | How to read it |
|---|---|---|
| Annual average LBMA PM price | $3,431.5 per troy ounce | The average across the year; up 44% from the 2024 annual average. |
| Year-end LBMA PM price | $4,368 per troy ounce | A year-end observation, not the annual average. |
| 2025 USD benchmark return | 67.4% | The full-year change reported in the December commentary table. |
Sources: World Gold Council, Gold Demand Trends: full-year 2025 (January 29, 2026), and December 2025 market commentary (January 8, 2026). The Council identifies ICE Benchmark Administration and other data providers in its reports. These are limited extracts for review and commentary.
Why 44% and 67.4% are different
An annual-average comparison and a year-end return use different observations. A forecast of an annual average should be compared with the annual average. A forecast of the closing price needs the stated closing benchmark. Selecting whichever number looks more favorable does not measure prediction accuracy.
A checklist for reviewing an old forecast
- Preserve the original: keep the publication date, exact prediction and source. A later rewrite cannot prove what was predicted earlier.
- Identify the measure: price target, annual average, return range or directional view.
- Match the units: US dollars per troy ounce differs from another currency or a price per gram.
- Match the deadline: an intrayear target and a December closing forecast are different tests.
- Record the outcome and miss: use the same benchmark and show the calculation rather than labelling a forecast “accurate” without it.
We are not presenting a league table of 2025 forecasters here. That would require the original dated forecasts and a consistent comparison method.
A benchmark return is not a buyer’s net return
A physical gold buyer pays a purchase price and receives a resale quote, with any storage, insurance and transaction charges in between. A fund investor has the fund’s own structure and expenses. An IRA adds account and custody terms. None is automatically identical to the benchmark figure.
Hypothetical cost example: buying an item for $10,000 and immediately receiving a $9,200 net resale quote creates an $800 gap, or 8% of the purchase price. This example is arithmetic, not a quote from a company or an estimate of a future return.
Use the CFTC and FINRA purchase checklist to identify the questions to put in a written offer.
Using the history today
Past price changes do not establish the next year’s outcome. Keep historical evidence separate from a current scenario or investment decision. For a forward-looking discussion, read our 2026 gold outlook. For a purchase decision, start with ways to invest in gold and the IRA cost worksheet.