Abstract BRAF ‐V600E mutation (mt) is a strong negative prognostic and predictive biomarker in metastatic colorectal cancer (mCRC). Non‐V600Emt, designated atypical BRAF mt (a BRAF mt) are rare, and little is known about their frequency, co‐mutations and prognostic and predictive role. These were compared between mutational groups of mCRC patients collected from three Nordic population‐based or real‐world cohorts. Pathology of a BRAF mt was studied. The study included 1449 mCRC patients with 51 (3%) a BRAF mt, 182 (13%) BRAF‐ V600Emt, 456 (31%) RAS & BRAF wild‐type (wt) and 760 (52%) RAS mt tumours. a BRAF mt were seen in 2% of real‐world and 4% of population‐based cohorts. Twenty‐six different a BRAF mt were detected, 11 (22%) class 2 (serrated adenocarcinoma in 2/9 tested), 32 (64%) class 3 (serrated in 15/25) and 4 (8%) unclassified. a BRAF mt patients were predominantly male, had more rectal primaries, less peritoneal metastases, deficient mismatch repair in one (2%), and better survival after metastasectomy (89% 5‐year overall survival [OS]‐rate) compared with BRAF ‐V600Emt. a BRAF mt and BRAF ‐V600Emt had poorer performance status and received fewer treatment lines than RAS & BRAF wt and RAS mt. OS among a BRAF mt (median 14.4 months) was longer than for BRAF‐ V600Emt (11.2 months), but shorter than for RAS & BRAF wt (30.5 months) and RAS mt (23.4 months). Addition of bevacizumab trended for better OS for the a BRAF mt. Nine patients with a BRAF mt received cetuximab/panitumumab without response. a BRAF mt represents a distinct subgroup differing from other RAS / BRAF groups, with serrated adenocarcinoma in only half. OS for patients with a BRAF mt tumours was slightly better than for BRAF ‐V600Emt, but worse than for RAS mt and RAS & BRAF wt. a BRAF mt should not be a contraindication for metastasectomy.