Building carbon intensity is related to material choice, but more importantly, material volume. The building structural frame itself is responsible for 20-30% of whole-life carbon over 50 years. This figure will double once we build net-zero operational carbon buildings. Carbon savings in the use of materials are therefore the key to reducing the environmental impact of buildings. Recent studies have shown that up to 40% of material in building structural frames could be successfully removed without a ffecting design code compliance. This unnecessary overdesign of buildings is in part due to a lack of structural optimisation, and acceptance by designers of conservative serviceability assumptions that represent the “low hanging fruit” of reducing embodied carbon in buildings. This paper examines steel frames buildings to determine the carbon savings that can be achieved for cross-section optimisation, as this is the most accessible form of optimisation, without changing the floor system and beam layout. For this purpose the Lightest Beam Method (LBM) was developed that studied non-composite universal beams (UB) members in buildings. Choosing the lightest section with the Eurocodes we can achieve 26.5% of steel savings by mass, with a half of beams governed by serviceability limit states (SLS). If deflection is calculated using variable loads, the proportion of beams governed by the SLS drops to 31.1% giving additional 2.2% mass savings. The highest steel savings of 34.5% can be achieved for lower natural frequency assumptions (3 Hz) and using the average rather than the characteristic steel yield strength. In this case the proportion of beams by mass governed by SLS drops to 19.7%. Based on available case studies it was found that 1/3 of steel in the frames could have been saved which represents 36% of initial embodied carbon or 5% of whole-life carbon for the building over 60 years.