Baseball results Friday 25 September 2026: AI predictions review

Friday 25 September 2026 produced 16 matches across MLB and NPB, and the 10-model TuringStats consensus finished the day 7 of 16 correct. It was a slate that punished conviction: several of the most heavily backed picks failed outright, while a handful of tighter calls landed.

The biggest surprise came at Yankee Stadium, where all 10 models backed New York to beat Baltimore and the Orioles instead ran out 10-2 winners. A second Yankees-Orioles meeting later in the day went the expected way, but the split underlined how quickly a short series can turn.

Where the AI got it right

Boston's double-header against the Cubs was the consensus at its most reliable. In the opener, all 10 models called a Boston Red Sox vs Chicago Cubs prediction for the home side, and Boston edged it 4-3. The models had the winner right even if the most common scoreline of 5-3 was two runs off.

The second game followed the same script. A unanimous Boston Red Sox vs Chicago Cubs prediction landed as Boston won 2-0, this time with the consensus score of 5-4 overstating the offence on both sides.

In Detroit, 70% of models backed the Tigers and they delivered an 8-7 win over Pittsburgh. The Detroit Tigers vs Pittsburgh Pirates prediction was one of the day's less certain calls and one of its more accurate ones.

Toronto's 6-5 win over Cincinnati was another unanimous call that held. All 10 models took the Blue Jays, with a most common score of 6-4 against an average confidence of 58%, and the Toronto Blue Jays vs Cincinnati Reds prediction came within a single run of the exact margin.

The Chicago White Sox and Milwaukee Brewers also justified unanimous support. The Chicago White Sox vs Colorado Rockies prediction was rewarded with a 6-1 home win, and the Milwaukee Brewers vs St.Louis Cardinals prediction came in at 8-4, again with the winner correct but the 6-4 consensus score short of the mark.

Where it got it wrong

The Yankees-Orioles opener was the day's clearest failure of confidence. All 10 models picked New York at an average confidence of 63%, with 5-3 the most common score. Baltimore won 10-2. The later meeting between the same teams went to New York 6-3, so the consensus was not wrong about the matchup so much as about the first game.

Seattle's 4-6 loss to the Angels was another unanimous miss. Every model backed the Mariners at an average confidence of 61% and a most common score of 4-2; the Angels scored six. The Seattle Mariners vs Los Angeles Angels prediction was among the most confident calls of the slate.

Philadelphia's 0-2 defeat to Tampa Bay carried the same 100% consensus, at 58% average confidence and a 5-3 most common score. The Philadelphia Phillies vs Tampa Bay Rays prediction never looked like producing the expected runs.

Washington's 7-6 win over the Mets was notable because the consensus score of 3-5 was the wrong way round entirely. All 10 models took New York at 60% average confidence; the Nationals won by one. The Washington Nationals vs New York Mets prediction joins the day's list of unanimous misses.

Atlanta's 0-3 loss at Miami was the highest-confidence error on the board. Every model backed the Braves at an average confidence of 70%, with 3-6 the most common score, and the Marlins won 3-0. Elsewhere, the Athletics fell 5-7 to Houston despite 90% support, the Dodgers lost 2-4 to San Diego on the same 90% share, and Kansas City's 9-12 defeat to Cleveland came with the consensus split at 50%.

The single NPB fixture went the same way. All 10 models backed Hanshin to beat Yokohama at 58% average confidence and a 4-5 score; the BayStars won 2-1.

Models of the day

Five models shared the best return of the day, each finishing 8 of 16: GPT-5.6 Luna, DeepSeek V4 Pro, MiMo V2.5 Pro, Gemini 3.8 Flash and GLM 5.3. No model in the field cleared the 50% mark by more than a single result.

Kimi K3 was the weakest performer at 6 of 16. A middle group of four — Claude Sonnet 5, Mistral Medium 3.5, Qwen 3.8 Max and Grok 4.6 — finished 7 of 16, matching the overall consensus record exactly.

Full results

• MLB: Seattle Mariners 4-6 Los Angeles Angels — consensus Seattle Mariners win (100%) WRONG
• MLB: Athletics 5-7 Houston Astros — consensus Athletics win (90%) WRONG
• MLB: Los Angeles Dodgers 2-4 San Diego Padres — consensus Los Angeles Dodgers win (90%) WRONG
• MLB: Boston Red Sox 4-3 Chicago Cubs — consensus Boston Red Sox win (100%) RIGHT
• MLB: New York Yankees 2-10 Baltimore Orioles — consensus New York Yankees win (100%) WRONG
• MLB: Boston Red Sox 2-0 Chicago Cubs — consensus Boston Red Sox win (100%) RIGHT
• MLB: Detroit Tigers 8-7 Pittsburgh Pirates — consensus Detroit Tigers win (70%) RIGHT
• MLB: Philadelphia Phillies 0-2 Tampa Bay Rays — consensus Philadelphia Phillies win (100%) WRONG
• MLB: Washington Nationals 7-6 New York Mets — consensus New York Mets win (100%) WRONG
• MLB: New York Yankees 6-3 Baltimore Orioles — consensus New York Yankees win (100%) RIGHT
• MLB: Toronto Blue Jays 6-5 Cincinnati Reds — consensus Toronto Blue Jays win (100%) RIGHT
• MLB: Miami Marlins 3-0 Atlanta Braves — consensus Atlanta Braves win (100%) WRONG
• MLB: Chicago White Sox 6-1 Colorado Rockies — consensus Chicago White Sox win (100%) RIGHT
• MLB: Kansas City Royals 9-12 Cleveland Guardians — consensus Kansas City Royals win (50%) WRONG
• MLB: Milwaukee Brewers 8-4 St.Louis Cardinals — consensus Milwaukee Brewers win (100%) RIGHT
• NPB: Yokohama BayStars 2-1 Hanshin Tigers — consensus Hanshin Tigers win (100%) WRONG

Saturday's card is already loaded. The full set of model calls, confidence levels and projected scorelines for today's fixtures is available on the today's baseball predictions page.

Next up: today's baseball predictions from the same panel, and the results archive with every graded call. Forecasts are informational only.