Background Both methadone- and buprenorphine-treated opioid-dependent patients frequently show cognitive deficits in attention, working memory, and verbal memory. However, no study has compared these patient groups with each other during early opioid substitution treatment (OST). Therefore, we investigated attention, working memory, and verbal memory of opioid-dependent patients within six weeks after the introduction of OST in a naturalistic setting and compared to those of healthy controls. Methods The sample included 16 methadone-, 17 buprenorphine/naloxone-treated patients, and 17 healthy controls matched for sex and age. In both groups buprenorphine was the main opioid of abuse during the recent month. Benzodiazepine codependence, recent use, and comedication were also common in both patient groups. Analysis of variance was used to study the overall group effect in each cognitive test. Pair-wise group comparisons were made, when appropriate Results Methadone-treated patients, as a group, had significantly slower simple reaction time (RT) compared to buprenorphine/naloxone-treated patients. In Go/NoGo RT methadone patients were significantly slower than controls. Both patient groups were significantly debilitated compared to controls in working memory and verbal list learning. Only methadone patients were inferior to controls in story recall. In simple RT and delayed story recall buprenorphine/naloxone patients with current benzodiazepine medication (n = 13) were superior to methadone patients with current benzodiazepine medication (n = 13). When methadone patients were divided into two groups according to their mean dose, the patient group with a low dose (mean 40 mg, n = 8) showed significantly faster simple RT than the high dose group (mean 67 mg, n = 8). Conclusion Deficits in attention may only be present in methadone-treated early phase OST patients and may be dose-dependent. Working memory deficit is common in both patient groups. Verbal memory deficit may be more pronounced in methadone-treated patients than in buprenorphine/naloxone-treated patients. In sum, to preserve cognitive function in early OST, the use of buprenorphine/naloxone may be more preferable to methadone use of, at least if buprenorphine has been recently abused and when benzodiazepine comedication is used. Longitudinal studies are needed to investigate if the better performance of buprenorphine/naloxone-treated patients is a relatively permanent effect or reflects "only" transient opioid switching effect.