Background . Several induction quality scoring systems (IQSS) have been described to evaluate drugs and risk factors of this perilous anaesthetic period in horses, but no attempts to compare their reliability have been performed. Objectives . To elucidate reliability of three IQSS: the visual analogue scale (VAS), a simple descriptive scale (SDS) and a composite grading scale (CGS) proposed by the authors. Study design. Reliability study. Methods. Eight randomly selected video-recorded anaesthetic inductions from horses that underwent general anaesthesia were evaluated by 4 observers twice. The three aforementioned IQSS were used. To assess reliability, intra and inter-rater intraclass correlation coefficient (ICC), and their 95%, confidence intervals (CI) were calculated based on a mean-rating (k = 4), absolute-agreement, 2-way random-effects model. Results . The higher inter-rater ICC found was for the VAS, presenting almost a very good reliability (0.74 ± 0.11). The CGS and the SDS had lower inter-rater agreement than the VAS, classified as moderate to good reliability (0.65 ± 0.22 and 0.63 ± 0.21, respectively). Intra-rater agreement results demonstrated very good reliability for both VAS and SDS (0.82 ± 0.08; 0.81 ± 0.18, respectively); and excellent reliability for the CGS (0.91 ± 0.08). Main limitations . Limitations of this study include the use of video-recordings, that may affect inter-observer results when anaesthetic inductions are assessed in situ. Additionally, these findings are applicable only when free inductions are assessed. Conclusions . VAS and the proposed CGS are reliable IQSS in horses, outperforming the widely used SDS. For research purposes, the VAS is advised if multiple evaluators assess induction quality, whereas the CGS would be selected for studies involving a single observer. Standardizing the VAS as IQSS may facilitate broader multi-centre studies as it is a simple, user-friendly and adequate for short-lasting events scale.