I would agree with tigerle on general principles. How that looks in reality is quite a bit more complex, of course. Perhaps it is worthwhile to think about a few of the concepts and models that feed into inclusion, such as:
1. Universal Design for Learning (UDL). Embedding scaffolding and accommodations into instruction for all students. On the disability end of the curve, these allow students to receive content instruction, and demonstrate learning, at the highest level concepts and skills of which they are capable, with supports for deficit or asynchronous areas that are not essential to the content learning objectives.
An example of this would be a student enrolled in a high school biology class whose conceptual skills were appropriate to that level, but whose reading or writing skills were not. Such a student might be allowed to use text-to-speech and speech-to-text to access text and produce written responses. Or use alternate assignments to demonstrate mastery of the standards, such as an audio-visual or performance product, rather than written work.
The variant of this known as differentiated instruction is usually focused on curricular adaptations, typically within a narrow range. The absence of greater adaptation upwards reflects the origins of UDL and DI in special education, for the purpose of giving students with disabilities and English language learners access to grade-level work.
2. Age-matched socialization and social skill development. This is a high value in the literature on inclusion. As it happens, I don't personally find it be as high of a value, and would argue that a higher goal would be social skill development that encompasses relationships with diversity in age as well as ability. When we remove the age-locked element, it opens up additional models for inclusion, such as age-decoupled instructional groupings based on current instructional needs. This brings strategies such as SSA and double- (or more) promotion into the umbrella of inclusion, as inclusive in age, but matched in academic needs. With respect to some of the social concerns, one solution that integrates both social developmental needs and intellectual/academic needs might be multi-age/multi-grade clusters, with students sorted by instructional needs in each content area, regardless of nominal grade placement.
My own educational history includes being the precipitating case for promoting change in the private secondary school I attended. Years after I left the school, I returned for prepractice observation, and was pleased to see that they had maintained and extended into policy many of the practices that initially had been proposed by my parents as exceptions for my sake, such as placing students in individual courses based on their specific instructional needs, both above and below nominal grade placement. It had become a common enough occurrence that it was no longer noteworthy to most students or faculty.
3. The 1%. By which I mean the top and bottom 1%. Well, it's probably more like the top and bottom 1-3%. The US and Canadian systems actually have a decent amount of success or potential success with +/- 2 SD (which is the middle 92% or so). Another couple percent on each end can potentially have their needs met with hybrid solutions/partial inclusion (some inclusion, some special programming). For the last few percent, it becomes more challenging.
I am a huge proponent of inclusive education for the disability end of the exceptional population, as the ability to navigate the general population on a social and practical level is a much more important long-term skill for them than achieving age-appropriate academics is. But even there, the intensive instruction necessary to master life skills and functional academics, and the limited incremental benefit from being included in academic (out of their instructional zones) settings with NT age-peers, suggest to me that a better use of their K-12+ education years (to age 22, in the US) would limit time in inclusive educational settings to those with the most potential for community transfer, such as in non-core classes with greater social interaction (e.g., chorus, athletics, community service).
For the right-hand tail, I believe a similar phenomenon occurs. At some point, the incremental benefit of leaving learners with extremely advanced academic ability in more-or-less age-matched groups for the purpose of developing social and adaptive skills within their age cohort ought to be balanced with the effective use of instructional time to meet their academic needs. Inclusive (age-matched) education for those students also might be better applied judiciously, and principally to settings in which academics are not the focus, and which have more intrinsic social opportunities, such as the examples named above. An additional factor is that there is a marked subset of intellectually-advanced learners whose social-emotional development is also advanced, for whom age-locked placement does not advance any of their skill development (other than tolerance for those with less skill, which, while a highly worthwhile goal, and likely one with significant life-long value to these learners in particular, also is a pretty modest sole objective for 10-13 years of compulsory education).
I could say more, but I think I'll leave it at that for the moment!
