Norris and McLaren Master Interlagos Weather
Lando Norris claimed victory at the Brazilian Grand Prix, capitalising on a fresh technical package to defeat Max Verstappen and Isack Hadjar in wet conditions at Autódromo José Carlos Pace. Heading into Round 20 of the 2026 season, McLaren sat fourth in the constructors’ standings on 363 points, trailing third-placed Red Bull by 16 points. After Norris finished fifth in Mexico City roughly 12 seconds behind Verstappen, [1] reported that McLaren would introduce upgrades to its front wing, rear wing, and floor components on the MCL40 for Brazil.
Those updates immediately paid dividends in a race defined by rain, with air temperatures at 16.1 °C and the track at 24.8 °C. Despite the wet surface, the race featured 20 classified finishers and zero safety cars, virtual safety cars, or red flags. Norris executed a three-stop strategy, starting with a 16-lap stint on intermediate tyres before switching to softs for stints of 33 and 13 laps, and returning to intermediates for the final 9 laps. Verstappen shadowed that approach—running 17 laps on intermediates, 28 and 17 on softs, and 9 on intermediates—to take second place after entering the weekend on the back of three wins in four races, as noted by [2]. Hadjar, who had been spun by George Russell in Mexico according to [3], rebounded on an identical tyre sequence to Norris (16 intermediate, 34 soft, 12 soft, 9 intermediate) to secure third.
Mercedes Outscores Ferrari Amid Midfield Penalties
Behind the podium trio, the battle between Mercedes and Ferrari carried major championship implications. Entering the weekend with 630 points to Ferrari’s 504, Mercedes needed to outscore their rivals by just three points in São Paulo to clinch the constructors’ title, according to [2]. They comfortably achieved that target as George Russell finished fourth and newly crowned world champion Andrea Kimi Antonelli took fifth, while Ferrari’s Charles Leclerc and Lewis Hamilton finished eighth and ninth.
Russell overcame a five-place grid penalty incurred for taking his sixth internal combustion engine of the season, a setback detailed by [4]. Unlike the three-stopping leaders, both Russell and Leclerc opted for two stops, running 16 laps on intermediates, 45 laps on a single set of softs, and 10 laps on intermediates. Antonelli stayed on a three-stop plan (17 intermediate, 36 soft, 9 soft, 9 intermediate) and recorded the fastest lap of the race with a 72.689-second tour on lap 55. Ferrari had held pre-race talks on Thursday after a first-corner clash between Hamilton and Leclerc in Mexico, with team boss Fred Vasseur stating the matter was resolved via [2], yet neither driver could crack the top seven.
Oscar Piastri brought the second McLaren home in sixth ahead of RB’s Liam Lawson in seventh. Nico Hülkenberg claimed tenth for Audi despite missing second practice with undisclosed car issues reported by [5] and serving a 10-second penalty for leaving the track and gaining an advantage. Across the 58 total pit stops—ranging from 23.3 seconds to 39.7 seconds—Williams driver Alexander Albon also served a 5-second penalty for rejoining unsafely.
AI Predictions: Norris Victory Defies the Models
The surprise victory for Norris and podium appearance for Hadjar caught every AI prediction model off guard, resulting in no “Model of the race” award being issued. All six models failed to call the winner and missed the exact podium set, heavily favouring pre-race championship pacesetters Antonelli and Verstappen instead.
Four of the six models—Claude Keeper, Kimi Cool, Gemini Thunder, and GPT Striker—backed Antonelli to win with Verstappen in second. Among that group, Gemini Thunder achieved the best overall probability calibration with a Brier score of 0.865, narrowly ahead of Claude Keeper at 0.8766 and Kimi Cool at 0.9502. All three of those models projected a podium of Antonelli, Verstappen, and Norris, meaning they correctly identified two of the three podium finishers, albeit with Norris in third rather than first and Antonelli finishing fifth.
DeepThinker and the Baseline model both predicted Verstappen to win ahead of Antonelli in second. DeepThinker placed Norris third and finished with the third-best Brier score at 0.9, while Baseline slotted Leclerc into third and recorded a Brier score of 0.9581. GPT Striker struggled the most among the cohort, pairing its Antonelli-Verstappen top two with a third-place call for Hamilton—who finished ninth—resulting in the highest Brier score of the weekend at 1.065.
Sources
- Autosport: McLaren to bring upgrades to Brazil GP with F1 2027 in mind motorsport.com ↩
- BBC Sport: Ferrari's Hamilton-Leclerc problem - and other F1 stories to follow bbc.co.uk ↩
- Autosport: Mercedes takes blame for Russell's poor Mexico GP: "We let him down" motorsport.com ↩
- Autosport: "Exhausted" Russell set for Brazilian F1 grid penalty motorsport.com ↩
- Autosport: F1 Brazilian GP: Antonelli quickest in tricky FP2 autosport.com ↩