Post-mortem staging of Alzheimer’s disease (AD) neurofibrillary pathology is commonly performed by immunohistochemistry using AT8 antibody for phosphorylated tau (p-tau) at positions 202/205. Thus, quantification of p-tau205 and p-tau202 in cerebrospinal fluid (CSF) should be more reflective of neurofibrillary tangles in AD than other p-tau epitopes. We developed two novel Simoa immunoassays for CSF p-tau205 and p-tau202 and measured these phosphorylations in three independent cohorts encompassing the AD continuum, non-AD cases and cognitively unimpaired participants: a discovery cohort (n = 47), an unselected clinical cohort (n = 212) and a research cohort well-characterized by fluid and imaging biomarkers (n = 262). CSF p-tau205 increased progressively across the AD continuum, while CSF p-tau202 was increased only in AD and amyloid (Aβ) and tau pathology positive (A+T+) cases (P < 0.01). In A+ cases, CSF p-tau205 and p-tau202 showed stronger associations with tau-PET (r_Sp205 = 0.67, r_Sp202 = 0.45) than Aβ-PET (r_Sp205 = 0.40, r_Sp202 = 0.09). CSF p-tau205 increased gradually across tau-PET Braak stages (P < 0.01), whereas p-tau202 only increased in Braak V–VI (P < 0.0001). Both showed stronger regional associations with tau-PET than with Aβ-PET, and CSF p-tau205 was significantly associated with Braak V–VI tau-PET regions. When assessing the contribution of Aβ and tau pathologies (indexed by PET) to CSF p-tau205 and p-tau202 variance, tau pathology was found to be the most prominent contributor in both cases (CSF p-tau205: R² = 69.7%; CSF p-tau202: R² = 85.6%) Both biomarkers associated with brain atrophy measurements globally (r_Sp205 = − 0.36, r_Sp202 = − 0.33) and regionally, and correlated with cognition (r_Sp205 = − 0.38/− 0.40, r_Sp202 = − 0.20/− 0.29). In conclusion, we report the first high-throughput CSF p-tau205 immunoassay for the in vivo quantification of tau pathology in AD, and a potentially cost-effective alternative to tau-PET in clinical settings and clinical trials.