Computer-based crop simulation models (CSMs) are well entrenched as tools for a wide variety of research, development and extension applications. Despite this, critics remain and there are perceptions that CSMs have not contributed to impacts on-farm or in the research community, particularly with plant breeding. This study reviewed the literature, interviewed 45 stakeholders (modellers, institutional representatives and clients of modelling), and analysed the industry-funded project portfolio to ascertain the current state of use of CSMs in the grains industry in Australia, including scientific progress, impacts and development needs. We found that CSMs in Australia are widely used, with ∼100 active and independent users, ∼15 model developers, and at any one time ∼10 postgraduate students, chiefly across six public research institutions. The dominant platform used is APSIM (Agricultural Production Systems Simulator). It is widely used in the agronomic domain. Several cases were documented where CSM use had a demonstrable impact on farm and research practice. The updating of both plant and soil process routines in the models has slowed and even stalled in recent years, and scientific limitations to future use were identified: the soil–plant nitrogen cycle, root growth and function, soil surface water and residue dynamics, impact of temperature extremes on plant function, and up-to-date cultivar parameter sets. There was a widespread appreciation of and optimism for the potential of CSMs to assist with plant-breeding activities, such as environmental characterisation, trait assessment, and design of plant-breeding programs. However, we found little evidence of models or model output being used by plant breeders in Australia, despite significant impacts that have emerged recently in larger international breeding programs. Closer cooperation between geneticists, physiologists and breeders will allow gene-based approaches to characterise and parameterise cultivars in CSMs, demonstrated by recent progress with phenology in wheat. This will give models the ability to deal with a wider range of potential genotype × environment × management scenarios.