Weather Models Forecast accuracy

Which weather forecast is most accurate where you are?

Nine weather models, scored on how their forecasts for your place actually turned out.

Choose a place

Forecast made
ahead
Over the last
Rank by

League table

How to read this table, and its limits

Last month in London, New York and Sydney

The same scoring for three cities, forecasts made 3 days ahead, ranked by temperature. Computed when this page was last published; choose a place above for live numbers.

London

Most accurate here, 3 days ahead: ECMWF IFS (Europe). Its forecast highs and lows were off by 0.8 °C on average. Too close to call with it: ICON, CMA GRAPES, JMA and UKMO. Bottom of the table: ARPEGE (1.3 °C).

#ModelTemperatureavg errorHighsavg errorLowsavg errorTends to runRainwet/dry right
1ECMWF IFSEurope0.8 °C0.6 °C1.0 °Cabout right90%
2ICONDWD, Germany1.0 °C0.8 °C1.2 °Cabout right93%
3ECMWF AIFSEurope, AI1.0 °C0.8 °C1.2 °C1.0° too cold93%
4CMA GRAPESChina1.0 °C0.9 °C1.2 °C0.3° too cold87%
5JMAJapan1.1 °C0.9 °C1.3 °Cabout right87%
6UKMOMet Office, UK1.1 °C0.8 °C1.4 °C0.6° too warm97%
7GEMECCC, Canada1.2 °C0.8 °C1.6 °C0.6° too cold80%
8GFSNOAA, US1.2 °C1.0 °C1.3 °C1.0° too warm93%
9ARPEGEMétéo-France1.3 °C1.0 °C1.5 °C1.0° too cold87%

Every model above is scored on the same 30 days.

Forecasts made 3 days ahead, checked against ERA5 for 30 days, 15 Oct – 13 Nov 2026; 12 of them wet (1 mm or more of rain or melted snow).

New York

Most accurate here, 3 days ahead: ECMWF AIFS (Europe, AI). Its forecast highs and lows were off by 1.0 °C on average. Too close to call with it: GEM and ECMWF IFS. Bottom of the table: JMA (2.3 °C).

#ModelTemperatureavg errorHighsavg errorLowsavg errorTends to runRainwet/dry right
1ECMWF AIFSEurope, AI1.0 °C0.9 °C1.1 °Cabout right100%
2GEMECCC, Canada1.0 °C1.0 °C1.0 °Cabout right90%
3ECMWF IFSEurope1.1 °C1.2 °C1.0 °Cabout right90%
4ICONDWD, Germany1.4 °C1.2 °C1.6 °C0.9° too warm97%
5ARPEGEMétéo-France1.5 °C1.5 °C1.4 °Cabout right93%
6UKMOMet Office, UK1.6 °C1.4 °C1.8 °C0.4° too warm93%
7GFSNOAA, US1.8 °C1.0 °C2.6 °C1.1° too warm90%
8CMA GRAPESChina2.0 °C2.4 °C1.5 °C1.7° too cold93%
9JMAJapan2.3 °C1.6 °C2.9 °C0.8° too warm97%

Every model above is scored on the same 30 days.

Forecasts made 3 days ahead, checked against ERA5 for 30 days, 14 Oct – 12 Nov 2026; 5 of them wet (1 mm or more of rain or melted snow).

Sydney

Most accurate here, 3 days ahead: GEM (ECCC, Canada). Its forecast highs and lows were off by 0.8 °C on average. Too close to call with it: ECMWF IFS, UKMO, JMA, ECMWF AIFS and ARPEGE. Bottom of the table: ICON (2.0 °C).

#ModelTemperatureavg errorHighsavg errorLowsavg errorTends to runRainwet/dry right
1GEMECCC, Canada0.8 °C0.8 °C0.9 °C0.5° too cold60%
2ECMWF IFSEurope0.9 °C0.7 °C1.0 °C0.7° too cold70%
3UKMOMet Office, UK0.9 °C1.1 °C0.6 °C0.3° too warm63%
4JMAJapan1.1 °C1.2 °C0.9 °Cabout right83%
5ECMWF AIFSEurope, AI1.1 °C1.5 °C0.6 °C0.5° too cold60%
6ARPEGEMétéo-France1.1 °C1.4 °C0.8 °Cabout right70%
7CMA GRAPESChina1.2 °C1.1 °C1.3 °C0.6° too cold57%
8GFSNOAA, US1.5 °C1.8 °C1.2 °C0.4° too warm67%
9ICONDWD, Germany2.0 °C2.2 °C1.8 °Cabout right63%

Every model above is scored on the same 30 days.

Forecasts made 3 days ahead, checked against ERA5 for 30 days, 15 Oct – 13 Nov 2026; 10 of them wet (1 mm or more of rain or melted snow).

Computed on 19 November 2026.

Questions

Which weather forecast is most accurate?

There is no single answer: it depends on where you are, the season, what you measure and how far ahead. Official global verification has long ranked ECMWF's IFS at or near the top for forecasts several days ahead, but at one particular place another model - often one that blends in a regional high-resolution model - can do better. The table on this page scores nine models for the place you choose.

Is the European model (ECMWF) more accurate than the American model (GFS)?

Averaged over the whole globe, ECMWF's IFS usually verifies better than NOAA's GFS for forecasts several days ahead. At a single place over a single month the order often flips, and the gap between them is frequently too small to call. Choose your place to see how they compare there.

How is the accuracy measured?

For every hour of every day in the period, each model's forecast made 1, 2, 3, 5 or 7 days before that hour is taken from Open-Meteo's archive of previous model runs and compared with ERA5, the Copernicus reanalysis of what actually happened. Open-Meteo keeps each hour from the run made that long before it, so one day's hours can come from consecutive runs rather than a single one. The score is the average error in the daily high and low temperature, plus how often the model correctly called the day wet or dry. Wet means 1 mm or more of precipitation: rain, plus any snow counted as melted water. A day is scored only when every one of its hours has both a forecast and a record.

Why does the table stop about a week ago?

ERA5, the record of what actually happened, is published about five days after the fact. Using a faster source would mean scoring the models against one of the models themselves, so the most recent days are left out until ERA5 covers them.

Why not use weather station readings?

This page runs entirely in your browser, with no server or database, and free station data with worldwide coverage is not available that way. ERA5 is available for every point on Earth and treats every place the same way. It is a reconstruction from observations on a grid of roughly 9 to 25 km, so it smooths out very local effects, and it is produced by ECMWF, which may slightly favour ECMWF's own model.

Is this an official measure of forecast accuracy?

No - it is a bit of fun. Weather services and forecast centres verify their models with many measures at once, such as typical and worst-case errors, skill compared with the seasonal average, hit and false-alarm rates, and scores for probability forecasts, against weather stations, radar and analyses, over years and whole regions. Different measures and different periods can put the models in a different order, so a model topping this table for one month at one place is not a verdict on which model is best.

Is anything sent or stored?

Nothing is tracked and there is no account. The coordinates of the place you choose go to Open-Meteo to fetch the forecasts and the ERA5 record, and anything you type in the search box goes to Photon to look the place up. Your choices are stored in this browser only.