Predicted Premier League table 2026/27: five AI season forecasts, side by side
Before the 2026/27 season, five frontier AI models — GPT-5.6, Claude Opus 4.8, Grok 4.5, Kimi K3 and GLM-5.2 — each predicted the complete final Premier League table, independently. 4 of 5 name Arsenal as champion; every model's Champions League and relegation calls are below, side by side, with the full predicted tables. Published 3 August 2026 and frozen since — graded here after the final matchday.
Published 3 August 2026, before matchday 1 — this forecast is frozen. It will not be edited; at the end of the season it is graded, right here.
Champion, Champions League places, relegation — every model's call
Each call is derived from that model's full predicted table: rank 1 is the champion pick, ranks 1–4 are the Champions League places, the bottom ranks are relegation. The table is the single source of truth — the calls cannot contradict it.
| Model | Champion | Champions League (top 4) | Relegation |
|---|---|---|---|
| ScoreGPTconsensus of the five | Arsenal | Arsenal · Manchester City · Liverpool · Chelsea | Ipswich · Hull City · Coventry |
| GPT-5.6 | Arsenal | Arsenal · Manchester City · Chelsea · Tottenham | Fulham · Coventry · Hull City |
| Claude Opus 4.8 | Arsenal | Arsenal · Manchester City · Liverpool · Chelsea | Ipswich · Coventry · Hull City |
| Grok 4.5 | Manchester City | Manchester City · Arsenal · Liverpool · Chelsea | Hull City · Sunderland · Coventry |
| Kimi K3 | Arsenal | Arsenal · Liverpool · Manchester City · Chelsea | Coventry · Ipswich · Hull City |
| GLM-5.2 | Arsenal | Arsenal · Manchester City · Liverpool · Chelsea | Ipswich · Hull City · Coventry |
ScoreGPT is not a sixth independent model — it is a consensus synthesized from the five models' tables after they answered.
What exactly is forecast — the frozen definitions
These definitions are printed so the forecast stays gradable: what counts was fixed on publication day, not decided after the fact.
- Champion: 1st place in the final table.
- Champions League places: positions 1–4 of the final table. A possible fifth berth via the European Performance Spots is not part of this forecast.
- Relegation: the bottom three places of the final table — 18, 19 and 20.
The full predicted tables, model by model
ScoreGPT — champion pick: Arsenalconsensus of the five
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Arsenal | 87 | +48 |
| 02 | Manchester City | 85 | +44 |
| 03 | Liverpool | 80 | +39 |
| 04 | Chelsea | 75 | +26 |
| 05 | Aston Villa | 66 | +15 |
| 06 | Newcastle | 65 | +17 |
| 07 | Manchester United | 64 | +11 |
| 08 | Tottenham | 63 | +11 |
| 09 | Brighton | 57 | +5 |
| 10 | Bournemouth | 54 | +2 |
| 11 | Crystal Palace | 51 | -2 |
| 12 | Nottingham Forest | 49 | -5 |
| 13 | Brentford | 48 | -5 |
| 14 | Fulham | 47 | -7 |
| 15 | Everton | 46 | -8 |
| 16 | Leeds | 41 | -14 |
| 17 | Sunderland | 37 | -17 |
| 18 | Ipswich | 35 | -22 |
| 19 | Hull City | 29 | -30 |
| 20 | Coventry | 28 | -33 |
On Arsenal: “Four of five models place Arsenal first after their 2025-26 title; squad continuity, elite defending and Arteta's settled system make them narrow favorites.” — ScoreGPT
GPT-5.6 — champion pick: Arsenal
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Arsenal | 86 | +48 |
| 02 | Manchester City | 82 | +43 |
| 03 | Chelsea | 78 | +35 |
| 04 | Tottenham | 73 | +28 |
| 05 | Manchester United | 70 | +24 |
| 06 | Liverpool | 67 | +20 |
| 07 | Aston Villa | 62 | +16 |
| 08 | Bournemouth | 58 | +10 |
| 09 | Brighton | 55 | +8 |
| 10 | Brentford | 52 | +4 |
| 11 | Newcastle | 49 | +1 |
| 12 | Everton | 47 | -2 |
| 13 | Crystal Palace | 45 | -4 |
| 14 | Leeds | 43 | -7 |
| 15 | Nottingham Forest | 41 | -10 |
| 16 | Sunderland | 39 | -13 |
| 17 | Ipswich | 37 | -16 |
| 18 | Fulham | 34 | -20 |
| 19 | Coventry | 31 | -25 |
| 20 | Hull City | 27 | -31 |
On Arsenal: “Arteta retains the division's most settled elite structure after a title-winning season, while Hincapie and Tzolis improve defensive cover and left-sided penetration. Gyokeres, Saka, Rice, Odegaard and a deep back line give Arsenal the best balance of ceiling and week-to-week reliability.” — GPT-5.6
Claude Opus 4.8 — champion pick: Arsenal
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Arsenal | 88 | +48 |
| 02 | Manchester City | 82 | +40 |
| 03 | Liverpool | 74 | +28 |
| 04 | Chelsea | 70 | +20 |
| 05 | Newcastle | 68 | +18 |
| 06 | Aston Villa | 66 | +14 |
| 07 | Manchester United | 64 | +12 |
| 08 | Tottenham | 58 | +5 |
| 09 | Brighton | 56 | +4 |
| 10 | Nottingham Forest | 54 | 0 |
| 11 | Bournemouth | 53 | -1 |
| 12 | Crystal Palace | 51 | -3 |
| 13 | Everton | 49 | -5 |
| 14 | Fulham | 47 | -6 |
| 15 | Brentford | 46 | -7 |
| 16 | Leeds | 44 | -10 |
| 17 | Sunderland | 42 | -14 |
| 18 | Ipswich | 38 | -20 |
| 19 | Coventry | 33 | -28 |
| 20 | Hull City | 26 | -38 |
On Arsenal: “Defending champions after ending a 22-year title drought, Arteta's squad retains its core and continues to add depth, giving them the platform to go back-to-back for the first time since the 1930s.” — Claude Opus 4.8
Grok 4.5 — champion pick: Manchester City
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Manchester City | 92 | +58 |
| 02 | Arsenal | 87 | +48 |
| 03 | Liverpool | 82 | +42 |
| 04 | Chelsea | 74 | +28 |
| 05 | Newcastle | 69 | +22 |
| 06 | Aston Villa | 65 | +15 |
| 07 | Tottenham | 62 | +12 |
| 08 | Manchester United | 58 | +5 |
| 09 | Brighton | 55 | +4 |
| 10 | Bournemouth | 52 | 0 |
| 11 | Crystal Palace | 48 | -5 |
| 12 | Fulham | 46 | -8 |
| 13 | Brentford | 44 | -12 |
| 14 | Nottingham Forest | 42 | -15 |
| 15 | Everton | 40 | -18 |
| 16 | Leeds | 38 | -22 |
| 17 | Ipswich | 36 | -28 |
| 18 | Hull City | 33 | -35 |
| 19 | Sunderland | 30 | -40 |
| 20 | Coventry | 26 | -48 |
On Manchester City: “Pep Guardiola's side retain elite squad depth and recruitment edge despite European demands; Haaland and new midfield additions drive another title charge.” — Grok 4.5
Kimi K3 — champion pick: Arsenal
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Arsenal | 88 | +48 |
| 02 | Liverpool | 85 | +45 |
| 03 | Manchester City | 83 | +44 |
| 04 | Chelsea | 74 | +25 |
| 05 | Manchester United | 70 | +15 |
| 06 | Newcastle | 68 | +18 |
| 07 | Aston Villa | 66 | +12 |
| 08 | Tottenham | 63 | +10 |
| 09 | Brighton | 59 | +5 |
| 10 | Nottingham Forest | 55 | +2 |
| 11 | Bournemouth | 52 | -2 |
| 12 | Crystal Palace | 50 | -3 |
| 13 | Fulham | 48 | -4 |
| 14 | Everton | 46 | -6 |
| 15 | Brentford | 44 | -8 |
| 16 | Leeds | 40 | -12 |
| 17 | Sunderland | 37 | -15 |
| 18 | Coventry | 34 | -20 |
| 19 | Ipswich | 30 | -25 |
| 20 | Hull City | 26 | -32 |
On Arsenal: “After multiple near-misses, Arteta's settled core of Saka, Rice, Saliba and Ødegaard enters its prime with improved attacking depth finally matching their elite defensive metrics. Fewer dropped points against low blocks tip the balance in their favour.” — Kimi K3
GLM-5.2 — champion pick: Arsenal
| # | Team | Proj. points | Proj. GD |
|---|---|---|---|
| 01 | Arsenal | 87 | +50 |
| 02 | Manchester City | 85 | +46 |
| 03 | Liverpool | 83 | +42 |
| 04 | Chelsea | 75 | +22 |
| 05 | Newcastle | 71 | +20 |
| 06 | Aston Villa | 69 | +16 |
| 07 | Tottenham | 66 | +14 |
| 08 | Manchester United | 64 | +6 |
| 09 | Brighton | 58 | +4 |
| 10 | Crystal Palace | 55 | +1 |
| 11 | Bournemouth | 53 | -3 |
| 12 | Brentford | 52 | -4 |
| 13 | Fulham | 50 | -6 |
| 14 | Nottingham Forest | 48 | -7 |
| 15 | Everton | 45 | -11 |
| 16 | Leeds | 40 | -16 |
| 17 | Sunderland | 37 | -19 |
| 18 | Ipswich | 34 | -23 |
| 19 | Hull City | 30 | -29 |
| 20 | Coventry | 26 | -36 |
On Arsenal: “Arteta’s project finally bears fruit as a matured core—Saka, Ødegaard, Saliba—peaks together. Surgical summer signings add the final cutting edge without disrupting chemistry.” — GLM-5.2
Projected points and goal difference exactly as each model returned them on publication day.
That is one line of one model's reasoning. Every club's reasoning, model by model — free to start in the ScoreGPT app.
Download the free appHow this forecast was made
On 3 August 2026, each of the five models was given the confirmed list of Premier League clubs for 2026/27 and asked for the complete final table — every position, with projected points and goal difference — using its own knowledge of summer transfers, squad depth and European commitments, with live web search enabled. Each model answered independently; none saw another model's table.
The model names are frozen as they were on publication day. ScoreGPT's live match-prediction roster evolves, but a dated forecast keeps the identity of whoever made it — a retired model is never relabelled as its successor.
After the final matchday, this page grades every call above against the real Premier League table — champion, the four Champions League places, and relegation, exactly as defined. Hits and misses stay side by side at this same URL.
The process is documented in the methodology and every model's history is on the accuracy leaderboard.
Frequently asked questions
Frequently asked
▸How was this Premier League season forecast created?
Five frontier AI models — GPT-5.6, Claude Opus 4.8, Grok 4.5, Kimi K3 and GLM-5.2 — were each given the confirmed 2026/27 Premier League club list and asked, independently and with live web search enabled, for the complete final table. The ScoreGPT row is a consensus synthesized from those five tables afterwards.
▸Will the forecast be updated during the season?
No. It was published on 3 August 2026 and is frozen — a forecast that can be quietly edited is not a forecast. At the end of the season it is graded on this page, and the misses stay visible next to the hits.
▸What exactly gets graded at the end of the season?
Three calls per model, exactly as defined on this page: the champion (1st place), the Champions League places (how many of the named four finish in the top 4, order-insensitive), and relegation. No composite score and no single accuracy number — the calls stand side by side.
▸Is the ScoreGPT row a sixth AI model?
No — it is the consensus of the five, synthesized from their tables after they answered. It is labelled that way wherever it appears, and it is graded the same way as the five.
▸Who will win the Premier League in 2026/27 according to AI?
4 of the 5 AI models pick Arsenal as 2026/27 Premier League champion; each model's full reasoning behind its table is in the ScoreGPT app. The pick is frozen and graded here against the real final table in May.
▸Who will be relegated from the Premier League in 2026/27 according to AI?
ScoreGPT's consensus table puts Ipswich, Hull City, Coventry in the direct relegation places. Each model's own relegation call is in the tables above — frozen, and graded against the real final table in May.
AI season forecasts for other leagues
Predictions are probabilistic information, not betting advice. 18+, please play responsibly.