Background: Triple-negative breast cancers (TNBCs) are heterogeneous in molecular drivers and immune traits. Our previous studies classified TNBCs into four subtypes (luminal androgen receptor, LAR; immunomodulatory, IM; basal-like immune-suppressed, BLIS; and mesenchymal-like, MES). The subtyping-based treatment strategy was proved clinically feasible, with promising efficacy (Cell Res 2021, NCT03805399) in patients with heavily pretreated metastatic TNBC (mTNBC). Here, we conducted FUTURE-SUPER, a multi-cohort, randomized trial to further evaluate the efficacy of subtyping-based therapy in the first-line treatment of mTNBC. Methods: Patients with untreated locally recurrent inoperable or mTNBC were enrolled. After categorization into 5 cohorts according to TNBC subtypes and genomic biomarkers, patients were randomly assigned (in a 1:1 ratio) to receive nab-paclitaxel (control group) or nab-paclitaxel plus subtyping-based treatment (subtyping-based group): pyrotinib (LAR-HER2^mut), everolimus (LAR-PI3K/AKT^mut), camrelizumab and famitinib (IM), bevacizumab (BLIS/MES-PI3K/AKT^WT), everolimus (MES-PI3K/AKT^mut). The primary endpoint was progression-free survival (PFS). Secondary endpoints included PFS in each subtyping-based cohort, objective response rate (ORR), overall survival (OS) and safety. Results: Between July 29, 2020, and October 17, 2022, 139 patients were enrolled and randomly assigned. With a median follow-up of 18.5 mo (range, 3.5 to 30.3), 101 (72.7%) events of disease progression or death occurred. Median PFS was significantly longer in the subtyping-based group than in the control (11.3 mo vs 5.8 mo; hazard ratio [HR], 0.38; 95% confidence interval [CI], 0.25 to 0.57; P<0.0001). Greater magnitude of PFS benefits were seen for the IM (median, 15.1 mo vs 6.5 mo; HR 0.37; 95% CI, 0.19 to 0.73), BLIS (median, 9.4 mo vs 3.9 mo; HR 0.41; 95% CI, 0.16 to 0.56) and LAR-PI3K/AKT^mut (median, 19.1 mo vs 8.4 mo; HR 0.42; 95% CI, 0.12 to 1.53) subtypes. The ORR (measurable-disease population) was higher in the subtyping-based group than in the control group (80.0% [48/60] vs. 44.8% [26/58]; odds ratio, 0.20; 95% CI, 0.09 to 0.46). OS data were not mature. Most adverse events were of grade 1 or 2. Grade 3 and 4 toxicities were as expected, with neutropenia more frequently reported in the subtyping-based group (56.5% vs 26.1%). No treatment-related deaths were recorded. Further high-throughput biomarker analyses are ongoing. Conclusions: Subtyping-based precision treatment significantly prolonged PFS versus nab-paclitaxel in the first-line treatment of advanced TNBC, with manageable toxicities. Our findings highlight the profound clinical benefits with treatment optimization according to molecular/immunological subtypes in TNBC, outlining a direction for further exploration. Clinical trial information: NCT04395989.