Assembly elections 2026 | Another pollster predicts 192 seats for BJP in Bengal, close fight between UDF, LDF in Kerala

Some poll predicting agencies say BJP may win close to or even bag over 100 out of 126 constituencies.

The Chanakya exit poll for Assam assembly election 2026 is out and it has predicted a landslide victory for the BJP. Out of 126 constituencies, the saffron party is predicted to win 102 seats (give or take 9 seats) whereas the Congress party could win 23 (± 9 seats), with the remaining seats going to others.

Axis My India has also predicted a sweep by the BJP with 88-100 seats, taking it easily past the majority mark of 64. People's Pulse, however, thinks the saffron party will not win more than 72 seats, whereas Matrize has predicted 85-95 seats for the BJP.

In Kerala, the United Democratic Front (UDF) led by Congress would win 78-90 seats, per Axis My India, followed by the CPI(M)-led Left Democratic Front or LDF, that is predicted to win 49-62 seats and the NDA (0-3 seats).

Chanakya predicts that the UDF will win 69 seats (± 9 seats), in a close contest with the LDF that could win 64 seats (± 9 seats). The BJP is likely to win seven seats (± 4 seats), per Chanakya.

In Tamil Nadu, Chanakya says DMK will emerge victorious with 125 seats (± 11) out of 234. Surprisingly, actor Vijay's TVK is likely to be second in the race with 63 seats (± 11), followed by AIADMK (45 ± 11 seats).

The BJP, per Chanakya, will also trample the ruling TMC in West Bengal, winning 192 seats (± 11) out of 294. The TMC, on the other hand, may win 100 seats (± 11) whereas the remaining will be bagged by others.

Meanwhile, Axis My India has said that it will not release the exit poll results for West Bengal this time citing voters' refusal to participate in the survey.

"We observed an exceptionally high non-response rate, with approximately 70% of approached voters declining to participate in the survey. While some degree of respondent hesitancy is factored into our sampling models, a refusal rate of this magnitude materially exceeds historical norms and introduces a high degree of non-response bias," the agency posted on X.