The presidential runoff, in one clear view. Official national counts, state-by-state results and the polling context behind the contest.
Flávio holds a narrow numerical polling edge. Polling error, turnout and late choices could change the outcome. This is a support estimate from 10 firms, not a win probability or a race call.
Read how this snapshot is calculatedLatest included release per firm · valid / normalized support
18 selected runoff polls from 10 firms. This is a checked editorial snapshot, not an exhaustive or automatically updating poll archive. Each row links to its publisher, report or named news source. Five final national waves released October 24 were added on October 25. AtlasIntel, Datafolha and Nexus were checked in original reports or commissioning-publisher material; Quaest and Futura remain explicitly news-reported. CNT/MDA October 24 is held: reviewed coverage disagrees on total sample size (2,002 versus 2,032). This is a selected collection, not all Brazilian polling.
| Pollster / source | Fieldwork | Sample / error | Flávio / PL | Lula / PT | Vote base / registration |
|---|---|---|---|---|---|
| AtlasIntel/Bloomberg (opens in a new tab)Random digital recruitment · Publisher checked | 18–23 OctReleased 24 Oct | 4,990±1 pp · 95% | 50.8% | 49.2% | Valid votesBR-06380/2026Original Atlas/Bloomberg report, methodology pp. 4–5 and valid-vote chart p. 8; report hosted by CNN Brasil. National adult sample; separate valid-vote question n not stated. |
| Datafolha (opens in a new tab)Face-to-face interviews · Publisher checked | 23–24 OctReleased 24 Oct | 4,050±2 pp · 95% | 51.0% | 49.0% | Valid votesBR-09540/2026Commissioning publisher Folha; 4,050 voters in 124 municipalities. Rounded valid-vote shares; original release also reports 50.89 / 49.11 unrounded. Separate valid-vote question base not stated. |
| Genial/Quaest (opens in a new tab)Collection mode not stated in reviewed release · News-reported | 23–24 OctReleased 24 Oct | 3,702±2 pp · 95% | 51.0% | 49.0% | Valid votesBR-09158/2026CNN Brasil reports the Grupo Globo-commissioned national wave, ages 16+. Original full report not obtained; kept labelled news-reported. Valid votes exclude blank/null; separate question base not supplied. |
| BTG/Nexus (opens in a new tab)Telephone (CATI) · Publisher checked | 23–24 OctReleased 24 Oct | 2,000±2.2 pp · 95% | 50.5% | 49.5% | Valid votesBR-08234/2026All-interviewee valid-vote scenario, ages 16+, all 27 states; 90% of interviews on Oct 24. Does not substitute the separate likely-voter model (52.4 / 47.6). Separate valid-vote base not stated. |
| Futura/100% Cidades (opens in a new tab)Collection mode not stated in reviewed release · News-reported | 23–24 OctReleased 24 Oct | 2,000±2.2 pp · 95% | 50.1% | 49.9% | Valid votesBR-09726/2026CNN Brasil report; self-funded national wave. Original full report not obtained; kept labelled news-reported. Separate valid-vote question base not supplied. |
| PoderData/Aya (opens in a new tab)Telephone · Publisher checked | 19–21 OctReleased 22 Oct | 4,000±1.6 pp · 95% | 51.0% | 49.0% | Valid votesBR-07926/2026 |
| Gerp (opens in a new tab)Telephone (GerpCATI) · News-reported | 19–21 OctReleased 21 Oct | 2,400±2 pp · 95% | 53.0% | 47.0% | Valid votesBR-03835/2026 |
| Genial/Quaest (opens in a new tab)Face-to-face household interviews · News-reported | 17–20 OctReleased 21 Oct | 2,154±2 pp · 95% | 45.0% | 44.0% | All respondentsNormalized: 50.56 / 49.44BR-08535/2026 |
| Meio/Ideia (opens in a new tab)Telephone · News-reported | 15–19 OctReleased 21 Oct | 1,500±2.5 pp · 95% | 51.1% | 48.9% | Valid votesBR-08251/2026 |
| RealTime Big Data (opens in a new tab)Telephone · News-reported | 15–19 OctReleased 20 Oct | 2,000±2 pp · 95% | 52.0% | 48.0% | Valid votesBR-02952/2026 |
| PoderData/Aya (opens in a new tab)Telephone · Publisher checked | 16–18 OctReleased 19 Oct | 3,000±1.8 pp · 95% | 52.0% | 48.0% | Valid votesBR-02586/2026 |
| AtlasIntel/Bloomberg (opens in a new tab)Random digital recruitment · Publisher checked | 13–18 OctReleased 19 Oct | 4,998±1 pp · 95% | 51.1% | 48.9% | Valid votesBR-02418/2026 |
| BTG/Nexus (opens in a new tab)Telephone · Publisher checked | 16–18 OctReleased 19 Oct | 2,002±2 pp · 95% | 50.3% | 49.7% | Valid votesBR-04499/2026 |
| PoderData/Aya (opens in a new tab)Telephone · Publisher checked | 13–15 OctReleased 16 Oct | 3,000±1.8 pp · 95% | 50.0% | 50.0% | Valid votesBR-04258/2026 |
| CNT/MDA (opens in a new tab)Face-to-face interviews · News-reported | 13–15 OctReleased 16 Oct | 2,002±2.2 pp · 95% | 50.5% | 49.5% | Valid votesBR-00652/2026 |
| Datafolha (opens in a new tab)Face-to-face interviews · Publisher checked | 13–14 OctReleased 15 Oct | 2,520±2 pp · 95% | 51.0% | 49.0% | Valid votesBR-00466/2026 |
| Genial/Quaest (opens in a new tab)Face-to-face household interviews · News-reported | 10–13 OctReleased 14 Oct | 2,154±2 pp · 95% | 46.0% | 44.0% | All respondentsNormalized: 51.11 / 48.89BR-08729/2026 |
| AtlasIntel/Bloomberg (opens in a new tab)Random digital recruitment · Publisher checked | 03–08 OctReleased 09 Oct | 5,026±1 pp · 95% | 52.8% | 47.2% | Valid votesBR-03663/2026Fieldwork began before the first-round vote. Excluded from the current polling snapshot. |
Current snapshot: latest national runoff poll per firm, fieldwork starting after October 4 and ending within 14 days of October 24, 2026. Weight = square root of sample size (capped at 3,000) × recency decay with a seven-day half-life. Total-response polls are normalized to the two finalists; their published figures and original base stay visible in the table. The Atlas October 3–8 release crosses the first round and is excluded from the current snapshot.
No house-effect, turnout or demographic adjustment is applied. The snapshot has no calibrated confidence interval or win probability. Individual poll margins of error are reproduced as published and do not describe the uncertainty of this aggregate. Shared polling errors remain possible. Registration references identify releases; registration by TSE is not an endorsement of their estimates. Polling support is never used to colour the official results map.
PL · 22
PT · 13Elected status recorded by TSE. Completed runoff results.
Completed source snapshot. Open the result studio to replay the saved count observations.
Completed TSE runoff result archive. Live collection is off; first-round history remains available.
Colour shows each state’s leading candidate.
Official runoff votes only. Select a state for its tally; grey means no accepted votes. Every valid vote contributes to the national total. IBGE 2025 boundaries.
Runoff counts only · grey cards are awaiting votes. State leads are not seats won.
| State | First-place candidate | Margin | Vote gap |
|---|---|---|---|
| Lula | 0.05 pp | 225 | |
| Lula | 3.24 pp | 70,232 | |
| Flávio Bolsonaro | 4.91 pp | 588,612 | |
| Lula | 5.41 pp | 263,280 | |
| Flávio Bolsonaro | 7.02 pp | 65,312 | |
| Flávio Bolsonaro | 13.20 pp | 233,989 | |
| Flávio Bolsonaro | 13.60 pp | 1,273,823 | |
| Flávio Bolsonaro | 13.73 pp | 3,416,610 | |
| Lula | 14.28 pp | 259,741 | |
| Flávio Bolsonaro | 17.02 pp | 383,166 | |
| Flávio Bolsonaro | 19.91 pp | 1,279,024 | |
| Flávio Bolsonaro | 22.55 pp | 863,008 | |
| Flávio Bolsonaro | 23.92 pp | 356,496 | |
| Lula | 24.98 pp | 517,470 | |
| Lula | 28.23 pp | 709,750 | |
| Flávio Bolsonaro | 28.71 pp | 1,890,748 | |
| Lula | 32.01 pp | 1,798,413 | |
| Lula | 32.12 pp | 438,180 | |
| Lula | 32.42 pp | 1,815,928 | |
| Lula | 33.10 pp | 1,327,261 | |
| Flávio Bolsonaro | 35.82 pp | 168,037 | |
| Flávio Bolsonaro | 35.97 pp | 716,900 | |
| Lula | 37.64 pp | 3,222,186 | |
| Flávio Bolsonaro | 41.56 pp | 402,321 | |
| Flávio Bolsonaro | 41.61 pp | 1,872,852 | |
| Lula | 46.90 pp | 1,006,682 | |
| Flávio Bolsonaro | 48.20 pp | 156,902 |
9,316,747 first-round valid votes went to candidates who are now out. Explore how transfers and finalist turnout could change the arithmetic.
YOUR SCENARIO · NOT A FORECASTRetained finalist voters keep their candidate. Unretained voters abstain or cast no valid vote; they do not switch sides. Returning eliminated-candidate voters split according to your slider. No new voters are added. This arithmetic cannot predict individual behaviour.
TSE official production JSON, presidential elections 6257 (first round) and 6258 (runoff). The completed runoff is preserved as a dated result archive. Automatic refresh and live collection are off. National and state records retain their original source times.
TSE technical documentation (opens in a new tab)Only official-environment files are accepted. Runoff votes dated before October 25 at 17:00 BRT are rejected. Test and first-round data cannot populate the runoff. An elected label requires TSE’s own candidate status.
Map: IBGE 2025 state and Federal District boundaries, simplified for display. First-round national totals reconcile with all 27 geographic units plus overseas voters. Finalist photographs come directly from TSE; party logos are the PT and PL marks, archived through Wikimedia Commons.
IBGE boundary files (opens in a new tab)PT logo provenance (opens in a new tab) · PL logo provenance (opens in a new tab)
Portraits identify the presidential candidates, not the administrative heads of their parties. Party branding does not imply endorsement. This desk covers the presidency; legislative and gubernatorial contests are not included.