Which weather model is most accurate in Boston?

Outlook Model accuracy

Boston, Massachusetts · daily high temperature in °C · trailing 30 days · refreshed at most every 6 hours

Every day, The Spread archives what each major weather model predicted for Boston at 1-, 3-, and 7-day leads, then scores those forecasts against the BOS observations once the day has happened. This page covers the trailing 30 days (2026-08-04 to 2026-09-02), for daily high temperature.

One day out, too close to call; at three days, too close to call; at seven days, too close to call.

The largest standing bias belongs to GEM (Canada), which has run cool by about 2.9°C at 3-day lead over this window.

The Spread's own bias-corrected blend scored a 0.82°C typical miss at one day. Its corrections come from the same window it is scored on (in-sample), so read it as a ceiling, not a promise.

The scoreboard

Typical miss (MAE) in °C with the signed bias in parentheses — positive bias means the model ran warm. Days = scored days behind the 1-day number.

Model	1-day	3-day	7-day	Days
DWD ICON	1.0° (+0.1°)	1.5° (+0.4°)	—	30
UKMO (UK)	1.0° (+0.3°)	1.4° (+0.3°)	—	30
GEM (Canada)	1.1° (+1.1°)	3.2° (-2.9°)	3.1° (-2.3°)	30
ECMWF IFS	1.2° (+0.0°)	1.5° (+0.1°)	3.0° (-0.1°)	30
NOAA GFS	1.2° (-1.0°)	1.7° (-1.3°)	2.6° (-0.2°)	30
NOAA AIGFS (AI)	1.4° (-0.1°)	1.6° (-0.0°)	3.5° (+0.2°)	30
ECMWF AIFS (AI)	1.6° (-1.2°)	1.8° (-1.3°)	2.5° (-1.0°)	30
NBM (US blend)	1.7° (-1.4°)	2.0° (-1.3°)	2.9° (-1.3°)	30
Blended (bias-corrected)	0.8° (-0.0°)	1.3° (-0.0°)	2.5° (+0.0°)	30

Method, honestly

The Spread's Boston point is the Boston Logan station (BOS), 4.1 km from downtown, used for both forecasts and verifications. Forecasts come from each model's actual archived runs via Open-Meteo's Previous Runs API — what the model said at the time, not a reconstruction. Truth is BOS observations — the station's own daily highs, a thermometer rather than a grid, and no correction applies: the thermometer is the truth outright. Rankings cover the models themselves; the two consensus products at the bottom of the table — NBM, which is itself a blend of models, and The Spread's own bias-corrected blend — are shown for comparison, never ranked or named "most accurate" here. The blend's corrections come from the same trailing window it is scored on (in-sample), so treat its score as a ceiling. A rolling 30-day window means these numbers move — that is the point.

Receipts — was the forecast right?

One page per settled date: what every model said as the day approached, what happened, and who was closest.

Tue Sep 1 12 days out the models said 21–32; it hit 21. GEM led (off 0.0°C); UKMO ran 1°C hot.
Mon Aug 31 11 days out the models said 20–31; it hit 23. UKMO led (off 0.6°C); AIGFS ran 3°C hot.
Sun Aug 30 10 days out the models said 23–31; it hit 30. ECMWF led (off 0.1°C); AIFS ran 3°C cold.
Sat Aug 29 9 days out the models said 24–28; it hit 23. GFS led (off 0.1°C); ECMWF ran 3°C hot.
Fri Aug 28 8 days out the models said 24–27; it hit 28. GFS led (off 0.0°C); ECMWF ran 2°C hot.
Thu Aug 27 7 days out the models said 20–26; it hit 29. GEM led (off 0.4°C); AIFS ran 3°C cold.
Wed Aug 26 6 days out the models said 23–28; it hit 26. AIFS led (off 0.4°C); AIGFS ran 2°C hot.
Tue Aug 25 5 days out the models said 25–27; it hit 27. GEM led (off 0.7°C); AIFS ran 2°C cold.
Mon Aug 24 4 days out the models said 25–29; it hit 29. GEM led (off 0.4°C); AIFS ran 3°C cold.
Sun Aug 23 3 days out the models said 21–28; it hit 22. GEM led (off 0.7°C); UKMO ran 5°C hot.
Mon Aug 17 13 days out the models said 20–30; it hit 24. ECMWF led (off 0.2°C); AIGFS ran 3°C cold.
Sun Aug 16 13 days out the models said 20–30; it hit 26. GFS led (off 0.1°C); UKMO ran 2°C hot.
Sat Aug 15 13 days out the models said 24–27; it hit 26. AIGFS led (off 0.1°C); GEM ran 2°C hot.
Fri Aug 14 13 days out the models said 18–27; it hit 28. ECMWF led (off 0.2°C); GFS ran 2°C hot.
Thu Aug 13 13 days out the models said 20–29; it hit 28. ICON led (off 0.1°C); ECMWF ran 1°C cold.
Wed Aug 12 13 days out the models said 22–27; it hit 29. AIGFS led (off 0.1°C); GEM ran 2°C hot.
Tue Aug 11 13 days out the models said 19–32; it hit 30. UKMO led (off 0.2°C); AIGFS ran 1°C hot.
Mon Aug 10 13 days out the models said 25–34; it hit 29. ICON led (off 0.2°C); ECMWF ran 2°C hot.
Sun Aug 9 13 days out the models said 25–29; it hit 33. GFS led (off 0.1°C); AIFS ran 2°C cold.
Sat Aug 8 12 days out the models said 28–30; it hit 34. GEM led (off 0.8°C); AIFS ran 3°C cold.
Fri Aug 7 11 days out the models said 27–31; it hit 35. GFS led (off 0.4°C); AIFS ran 5°C cold.
Thu Aug 6 10 days out the models said 23–32; it hit 32. GEM led (off 0.1°C); AIFS ran 2°C cold.
Wed Aug 5 9 days out the models said 26–32; it hit 31. ICON led (off 0.2°C); AIFS ran 2°C cold.
Tue Aug 4 8 days out the models said 26–28; it hit 29. AIGFS led (off 0.0°C); UKMO ran 2°C hot.
Mon Aug 3 7 days out the models said 21–27; it hit 25. AIGFS led (off 0.4°C); UKMO ran 2°C hot.
Sun Aug 2 6 days out the models said 20–27; it hit 29. ECMWF led (off 0.3°C); GEM ran 2°C hot.
Sat Aug 1 5 days out the models said 20–30; it hit 30. GFS led (off 0.1°C); AIFS ran 2°C cold.
Fri Jul 31 4 days out the models said 20–29; it hit 25. ICON led (off 0.1°C); GFS ran 4°C hot.
Thu Jul 30 3 days out the models said 19–23; it hit 23. GEM led (off 1.4°C); AIGFS ran 4°C cold.

Other cities

Atlanta, Georgia Berlin, State of Berlin Buffalo, New York Charlotte, North Carolina Chicago, Illinois Dallas, Texas Denver, Colorado Detroit, Michigan Houston, Texas London, England Los Angeles, California Miami, Florida Milan, Lombardy Minneapolis, Minnesota New Orleans, Louisiana New York, New York Oklahoma City, Oklahoma Orlando, Florida Paris, Île-de-France Region Philadelphia, Pennsylvania Phoenix, Arizona Salt Lake City, Utah San Antonio, Texas San Diego, California San Francisco, California Seattle, Washington St. Louis, Missouri Tampa, Florida Washington, District of Columbia

The Spread — the forecast spread — is in closed beta. These verification pages are public while the rest is gated.