PLAYSTAT / NFL MODEL COMPARISON

One slate. Three perspectives.

Compare GPT Terra, GPT Astra and DeepSeek across 12 NFL games. Open a matchup to see the evidence and counterargument behind each call.

9 November 2026 · all kickoff times Asia/Manila · inputs captured 8 Nov, 18:05 Manila

Published 8 November 2026. DeepSeek is the revised run after factual-grounding corrections; Terra and Astra retain their original estimates.

Dated research, not a live feed. Percentages are subjective, uncalibrated estimates conditional on a decisive result. All models saw ESPN’s baseline. No validated accuracy record; agreement does not establish an edge.

Results & model history ↗ Results will appear after verified finals.

NFL model estimates · 9 November 2026 · conditional on a decisive result
Matchup / Manila time	GPT Terra	GPT Astra	DeepSeek
Dallas Cowboys at Indianapolis Colts9 Nov · 02:00 Manila · Dated pregame snapshot	DAL 60%	DAL 60%	DAL 60%
New York Jets at Kansas City Chiefs9 Nov · 02:00 Manila · Dated pregame snapshot	KC 70%	KC 70%	KC 80%
Detroit Lions at Miami Dolphins9 Nov · 02:00 Manila · Dated pregame snapshot	DET 70%	DET 70%	DET 75%
Cleveland Browns at New Orleans Saints9 Nov · 02:00 Manila · Dated pregame snapshot	No pick	NO 55%	CLE 55%
New York Giants at Philadelphia Eagles9 Nov · 02:00 Manila · Dated pregame snapshot	NYG 60%	NYG 60%	NYG 65%
Los Angeles Rams at Washington Commanders9 Nov · 02:00 Manila · Dated pregame snapshot	LAR 65%	LAR 65%	LAR 70%
Denver Broncos at Carolina Panthers9 Nov · 02:00 Manila · Dated pregame snapshot	DEN 60%	DEN 55%	DEN 55%
Houston Texans at Los Angeles Chargers9 Nov · 05:05 Manila · Dated pregame snapshot	HOU 60%	HOU 65%	HOU 60%
Green Bay Packers at New England Patriots9 Nov · 05:25 Manila · Dated pregame snapshot	No pick	NE 55%	NE 55%
Las Vegas Raiders at San Francisco 49ers9 Nov · 05:25 Manila · Dated pregame snapshot	SF 70%	SF 75%	SF 80%
Arizona Cardinals at Seattle Seahawks9 Nov · 05:25 Manila · Dated pregame snapshot	SEA 65%	SEA 65%	SEA 70%
Tampa Bay Buccaneers at Chicago Bears9 Nov · 09:20 Manila · Dated pregame snapshot	CHI 60%	CHI 60%	CHI 60%

The reasoning, game by game.

Tap a matchup to expand

Dallas Cowboys at Indianapolis Colts9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

DAL 60%

medium confidence · subjective estimate

Evidence

Dallas scores 27.4 versus Indianapolis 20.1 points per game.
Dallas gains 349.9 versus 310.3 yards per game.

Counterargument

Dallas allows 396.9 versus Indianapolis 358.0 yards per game.

GPT AstraGPT-6 Astra

DAL 60%

low confidence · subjective estimate

Evidence

Points Per Game: IND 20.1; DAL 27.4.
Yards Passing: IND 193.8; DAL 255.8.

Counterargument

The Colts allow fewer total yards than Dallas, so the offensive edge does not establish a clean two-sided advantage.

DeepSeekDeepSeek V4 Pro

DAL 60%

medium confidence · subjective estimate

Evidence

Points per game favors DAL away 27.4 vs IND home 20.1
Passing yards per game favors DAL away 255.8 vs IND home 193.8

Counterargument

Rushing yards per game favors IND home 116.5 vs DAL away 94.1 and Points allowed per game is close at IND home 27.8 vs DAL away 26.6

Watched-wallet comparison loading…

ESPN baseline: DAL 61.1% · IND 38.6% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

New York Jets at Kansas City Chiefs9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

KC 70%

medium confidence · subjective estimate

Evidence

Kansas City gains 390.6 versus New York 344.3 yards per game.
Kansas City allows 18.7 versus New York 23.5 points per game.

Counterargument

New York allows 208.4 versus Kansas City 228.4 passing yards per game.

GPT AstraGPT-6 Astra

KC 70%

medium confidence · subjective estimate

Evidence

Total Yards: KC 390.6; NYJ 344.3.
Points Allowed Per Game: KC 18.7; NYJ 23.5.

Counterargument

The Jets allow fewer passing yards; Kansas City's passing-volume edge may not translate into an easy matchup.

DeepSeekDeepSeek V4 Pro

KC 80%

medium confidence · subjective estimate

Evidence

Points per game favors KC home 23.9 vs NYJ away 21.3
Points allowed per game favors KC home 18.7 vs NYJ away 23.5

Counterargument

Passing yards allowed per game favors NYJ away 208.4 vs KC home 228.4; NYJ passing offense is close at 239.1 vs KC 256.4

Watched-wallet comparison loading…

ESPN baseline: NYJ 19.7% · KC 80.2% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Detroit Lions at Miami Dolphins9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

DET 70%

medium confidence · subjective estimate

Evidence

Detroit scores 29.0 versus Miami 16.7 points per game.
Detroit gains 412.0 versus Miami 304.4 yards per game.

Counterargument

Detroit allows 395.3 versus Miami 344.0 yards per game.

GPT AstraGPT-6 Astra

DET 70%

medium confidence · subjective estimate

Evidence

Points Per Game: MIA 16.7; DET 29.0.
Total Yards: MIA 304.4; DET 412.0.

Counterargument

Detroit allows more passing yards than Miami; the weaker defense leaves an upset route despite Detroit's much stronger offensive figures.

DeepSeekDeepSeek V4 Pro

DET 75%

medium confidence · subjective estimate

Evidence

Points per game favors DET away 29.0 vs MIA home 16.7
Total yards per game favors DET away 412.0 vs MIA home 304.4

Counterargument

Yards allowed per game favors MIA home 344.0 vs DET away 395.3, limiting DET's total-yard edge

Watched-wallet comparison loading…

ESPN baseline: DET 77.6% · MIA 22.3% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Cleveland Browns at New Orleans Saints9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

No pick

low confidence · subjective estimate

Evidence

Cleveland allows 21.6 versus New Orleans 27.7 points per game.
Cleveland is 5-3 versus New Orleans 2-5.

Counterargument

New Orleans scores 23.1 versus Cleveland 17.4 points per game; ESPN favors New Orleans 68.0%.

GPT AstraGPT-6 Astra

NO 55%

low confidence · subjective estimate

Evidence

Points Per Game: NO 23.1; CLE 17.4.
Yards Passing: NO 245.9; CLE 226.0.

Counterargument

Cleveland allows fewer points; New Orleans's offensive edge conflicts with a much weaker scoring defense. ESPN is visible and helps support only a tentative New Orleans lean.

DeepSeekDeepSeek V4 Pro

CLE 55%

low confidence · subjective estimate

Evidence

Points allowed per game favors CLE away 21.6 vs NO home 27.7
Yards allowed per game favors CLE away 341.5 vs NO home 389.1

Counterargument

Points per game favors NO home 23.1 vs CLE away 17.4 and Passing yards per game favors NO home 245.9 vs CLE away 226.0

Watched-wallet comparison loading…

ESPN baseline: CLE 31.7% · NO 68.0% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

New York Giants at Philadelphia Eagles9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

NYG 60%

low confidence · subjective estimate

Evidence

New York scores 23.3 versus Philadelphia 17.1 points per game.
New York allows 320.3 versus Philadelphia 375.9 yards per game.

Counterargument

Philadelphia is at home and ESPN rates it 47.0% versus New York 52.7%.

GPT AstraGPT-6 Astra

NYG 60%

low confidence · subjective estimate

Evidence

Points Per Game: PHI 17.1; NYG 23.3.
Yards Allowed: PHI 375.9; NYG 320.3.

Counterargument

ESPN is close to a toss-up, and the packet lacks opponent-strength and sample-window checks that could explain the Giants's apparent statistical advantage.

DeepSeekDeepSeek V4 Pro

NYG 65%

medium confidence · subjective estimate

Evidence

Points per game favors NYG away 23.3 vs PHI home 17.1
Yards allowed per game favors NYG away 320.3 vs PHI home 375.9

Counterargument

Points allowed per game is close at PHI home 25.3 vs NYG away 23.0; NYG scoring edge is only 6.2 points per game

Watched-wallet comparison loading…

ESPN baseline: NYG 52.7% · PHI 47.0% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Los Angeles Rams at Washington Commanders9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

LAR 65%

medium confidence · subjective estimate

Evidence

Los Angeles gains 395.4 versus Washington 341.1 yards per game.
Los Angeles allows 21.0 versus Washington 26.0 points per game.

Counterargument

Washington allows 99.1 versus Los Angeles 109.3 rushing yards per game.

GPT AstraGPT-6 Astra

LAR 65%

medium confidence · subjective estimate

Evidence

Total Yards: WSH 341.1; LAR 395.4.
Points Allowed Per Game: WSH 26.0; LAR 21.0.

Counterargument

Washington allows fewer rushing yards; its run defense offers a counterweight to the Rams's broader offense and scoring-defense advantages.

DeepSeekDeepSeek V4 Pro

LAR 70%

medium confidence · subjective estimate

Evidence

Points per game favors LAR away 24.3 vs WSH home 22.7
Yards allowed per game favors LAR away 310.3 vs WSH home 359.1

Counterargument

Rushing yards allowed per game favors WSH home 99.1 vs LAR away 109.3; WSH rushingYardsPerGame 126.6 exceeds LAR away 125.5 by 1.1

Watched-wallet comparison loading…

ESPN baseline: LAR 72.9% · WSH 26.9% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Denver Broncos at Carolina Panthers9 Nov · 02:00 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

DEN 60%

low confidence · subjective estimate

Evidence

Denver is 6-2 versus Carolina 3-4.
Denver allows 20.6 versus Carolina 27.6 points per game.

Counterargument

Carolina scores 25.4 versus Denver 21.5 points per game.

GPT AstraGPT-6 Astra

DEN 55%

low confidence · subjective estimate

Evidence

Points Allowed Per Game: CAR 27.6; DEN 20.6.
Points Per Game: CAR 25.4; DEN 21.5.

Counterargument

Carolina produces far more total yards. Denver's lower points allowed and visible ESPN lean face a strong offensive countercase.

DeepSeekDeepSeek V4 Pro

DEN 55%

low confidence · subjective estimate

Evidence

Points allowed per game favors DEN away 20.6 vs CAR home 27.6
Rushing yards allowed per game favors DEN away 127.9 vs CAR home 140.7

Counterargument

Total yards per game favors CAR home 383.1 vs DEN away 279.3 and Passing yards per game favors CAR home 281.1 vs DEN away 186.6

Watched-wallet comparison loading…

ESPN baseline: DEN 63.2% · CAR 36.5% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Houston Texans at Los Angeles Chargers9 Nov · 05:05 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

HOU 60%

low confidence · subjective estimate

Evidence

Houston scores 23.6 versus Los Angeles 14.1 points per game.
Houston gains 351.4 versus Los Angeles 310.9 yards per game.

Counterargument

Los Angeles allows 332.3 versus Houston 360.7 yards per game.

GPT AstraGPT-6 Astra

HOU 65%

medium confidence · subjective estimate

Evidence

Points Per Game: LAC 14.1; HOU 23.6.
Total Yards: LAC 310.9; HOU 351.4.

Counterargument

The Chargers allow fewer passing yards; Houston's stronger passing production does not imply an uncontested matchup.

DeepSeekDeepSeek V4 Pro

HOU 60%

medium confidence · subjective estimate

Evidence

Points per game favors HOU away 23.6 vs LAC home 14.1
Passing yards per game favors HOU away 266.7 vs LAC home 199.7

Counterargument

Rushing yards per game favors LAC home 111.1 vs HOU away 84.7; Yards allowed per game favors LAC home 332.3 vs HOU away 360.7

Watched-wallet comparison loading…

ESPN baseline: HOU 61.9% · LAC 37.8% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Green Bay Packers at New England Patriots9 Nov · 05:25 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

No pick

low confidence · subjective estimate

Evidence

Both teams are 5-3.
New England allows 20.3 versus Green Bay 20.5 points per game.

Counterargument

New England gains 376.8 versus Green Bay 341.4 yards per game, and ESPN favors New England 56.6%.

GPT AstraGPT-6 Astra

NE 55%

low confidence · subjective estimate

Evidence

Total Yards: NE 376.8; GB 341.4.
Yards Rushing: NE 123.1; GB 60.6.

Counterargument

Green Bay produces more passing yards. The defenses have similar headline points-allowed figures, so New England's advantage remains narrow.

DeepSeekDeepSeek V4 Pro

NE 55%

low confidence · subjective estimate

Evidence

Points per game favors NE home 23.0 vs GB away 21.4
Rushing yards per game favors NE home 123.1 vs GB away 60.6

Counterargument

Passing yards per game favors GB away 280.8 vs NE home 253.6 and Passing yards allowed per game favors GB away 220.9 vs NE home 235.5

Watched-wallet comparison loading…

ESPN baseline: GB 43.1% · NE 56.6% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Las Vegas Raiders at San Francisco 49ers9 Nov · 05:25 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

SF 70%

medium confidence · subjective estimate

Evidence

San Francisco allows 19.9 versus Las Vegas 28.8 points per game.
San Francisco allows 312.9 versus Las Vegas 397.1 yards per game.

Counterargument

Las Vegas scores 27.6 points per game, close to San Francisco 28.1.

GPT AstraGPT-6 Astra

SF 75%

medium confidence · subjective estimate

Evidence

Points Allowed Per Game: SF 19.9; LV 28.8.
Yards Allowed: SF 312.9; LV 397.1.

Counterargument

Las Vegas produces nearly as many points as San Francisco; a strong opposing offense keeps this from being a high-confidence selection.

DeepSeekDeepSeek V4 Pro

SF 80%

medium confidence · subjective estimate

Evidence

Points allowed per game favors SF home 19.9 vs LV away 28.8
Yards allowed per game favors SF home 312.9 vs LV away 397.1

Counterargument

Passing yards per game favors LV away 273.1 vs SF home 249.7; both scoring offenses are close at SF 28.1 vs LV 27.6

Watched-wallet comparison loading…

ESPN baseline: LV 15.9% · SF 83.9% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Arizona Cardinals at Seattle Seahawks9 Nov · 05:25 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

SEA 65%

medium confidence · subjective estimate

Evidence

Seattle allows 19.9 versus Arizona 29.8 points per game.
Seattle allows 263.1 versus Arizona 359.8 yards per game.

Counterargument

The offenses gain almost the same total yards per game: Seattle 348.9, Arizona 348.5.

GPT AstraGPT-6 Astra

SEA 65%

medium confidence · subjective estimate

Evidence

Points Allowed Per Game: SEA 19.9; ARI 29.8.
Yards Allowed: SEA 263.1; ARI 359.8.

Counterargument

The two offenses have almost identical total-yard figures; Seattle needs its defensive edge to carry the selection.

DeepSeekDeepSeek V4 Pro

SEA 70%

medium confidence · subjective estimate

Evidence

Points allowed per game favors SEA home 19.9 vs ARI away 29.8
Yards allowed per game favors SEA home 263.1 vs ARI away 359.8

Counterargument

Passing yards per game is level at SEA home 239.3 and ARI away 239.3; Rushing yards per game is near level at SEA 109.6 vs ARI 109.3

Watched-wallet comparison loading…

ESPN baseline: ARI 33.2% · SEA 66.5% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

Tampa Bay Buccaneers at Chicago Bears9 Nov · 09:20 Manila · Open reasoning

GPT TerraGPT-6.1 Terra

CHI 60%

low confidence · subjective estimate

Evidence

Chicago gains 404.6 versus Tampa Bay 316.1 yards per game.
Chicago scores 26.8 versus Tampa Bay 22.3 points per game.

Counterargument

Tampa Bay allows 18.9 versus Chicago 20.0 points per game.

GPT AstraGPT-6 Astra

CHI 60%

low confidence · subjective estimate

Evidence

Points Per Game: CHI 26.8; TB 22.3.
Yards Passing: CHI 253.9; TB 147.5.

Counterargument

Tampa Bay allows fewer points, so Chicago's superior offensive production is counterbalanced by the visitors’ stronger scoring defense.

DeepSeekDeepSeek V4 Pro

CHI 60%

low confidence · subjective estimate

Evidence

Points per game favors CHI home 26.8 vs TB away 22.3
Total yards per game favors CHI home 404.6 vs TB away 316.1

Counterargument

Points allowed per game favors TB away 18.9 vs CHI home 20.0; Yards allowed per game favors TB away 300.1 vs CHI home 321.4

Watched-wallet comparison loading…

ESPN baseline: TB 37.1% · CHI 62.6% · draw unspecified. Source report ↗

Provider stats are pregame context with unverified sample windows. Final starters, inactive lists, weather and late news were not verified. Dated pregame snapshot.

How to read this edition.

Three separate model runs used the same frozen ESPN packet. Models did not see one another’s answers, but all could see ESPN’s estimates. The prompts and reasoning budgets were not identical. These are research opinions, not independently trained or calibrated sports models.

Source cut: 8 November 2026 at 18:05 Manila. A receipt time does not prove the underlying provider data is fresh. Stats have unverified sample windows and opponent adjustment. Final starters, inactive lists and weather were not verified.

DeepSeek review: the first draft incorrectly treated omitted interceptions as zero, inferred form chronology, used spread language without a spread, and made unsupported weather and league-wide claims. A separate revised run was checked against exact team statistics before publication. Its revised estimates replace that draft here; no percentages were edited by hand.

After kickoff: these remain dated pregame opinions and are marked archived. Verified final results and the original predictions are tracked separately in the model history. This edition does not claim validated predictive skill. An abstention means insufficient support for a pick, not zero probability.

ESPN home/away percentages are reproduced separately without normalization; an unspecified draw value is unknown. Model percentages are conditional on a decisive result, so precise differences from ESPN are not an established edge.

NFL Week 9 schedule ↗ · ESPN schedule ↗