Liverpool vs Manchester City prediction: 5 of 5 AI models back Manchester City

The read: a decimated Liverpool attack against a perfect City side

Every expert on the panel lands on the same headline, and the evidence behind it is unusually concrete rather than reputational. Liverpool's entire first-choice forward line is unavailable: Iraola confirmed on October 9 that Isak (thigh) and Gakpo (ankle) are out after injuries on international duty, joining Ekitike (Achilles) and the unused Chiesa on the absentee list. Isak alone had four of Liverpool's seven league goals. City, by contrast, have Haaland fit and confirmed ready by Maresca, and their only losses — Foden's final suspension game and the O'Reilly doubt — are materially less damaging than losing a strike force. This asymmetry, not history or atmosphere, is what decides the shape of the match.

Where the experts agree — and how far to trust it

• Form gap is real and recent: City are 5-from-5 with 13 goals; Liverpool are sixth, unbeaten but draw-heavy (W1 D4 in their last five home league games), scoring 1.4 per game with their forwards. Strip those forwards out and a realistic home scoring expectation drops toward a single goal — hence the near-unanimous 1-2 modal scoreline.
• Liverpool's floor is genuinely high: four conceded in five, the league's meanest defence, and an unbeaten record. GPT-6.1 Sol's tactical point is the sharpest of the panel — Liverpool lead the league in high-intensity pressing (48.9% of opposition touches) while City are the league's best at passing through it (83% completion under that pressure). That suggests Liverpool will create moments but not sustain control, which is exactly why the draw stays uncomfortably live.

Resolving the disagreements

The main fault line is over the verdict circus: several models weight the hostile Anfield reception following City's guilty finding as an emotional equaliser. Claude Opus 5.5 and GPT-6.1 Sol both treat it — plus Liverpool's four straight home draws and City's soft fixture run — as reason to cap City around 42-48% and prefer insurance. Grok, GLM and Kimi treat the personnel gap as the more decisive, more certain fact. The synthesis sides with the latter on direction but the former on calibration: an improvised Barcola/Wirtz/Koumas front line is the hardest fact in the fixture, yet a compact, well-drilled Liverpool block against a Foden-less City attack is a classic low-event script. That is why the panel's draw probability of 29% is respected rather than argued down, and why the 1-2 — Liverpool scoring once through a set piece, Van Dijk aerial or a Wirtz moment, City twice through Haaland and a wide runner — is the modal scoreline, with 1-1 the principal alternative.