Project Orion is detonating nuclear weapons which consist of far more than just fuel, they also need high explosives etc, however much of the fuels energy is wasted both from low yield bombs being less efficient but also simply radiating energy in all directions. Nuclear powered ION engines however would be ejecting far more of their reaction mass of say xenon per KG of fuel that leftover fuel is basically irrelevant. In other words it’s a more efficient nuclear reaction, a vastly more efficient engine, and a much lower mass fraction spent on the physical engine.
This gets more extreme when you consider a crewed mission would need lots of power during the trip so they would want/need to build redundant nuclear reactors anyway.
Wait, if that is true, then the ION engine has a lower ISP than direct explosions, not higher. Unless the reactor is orders of magnitude more efficient than the bombs.
Smaller nuclear detonations are less efficient in that they directly liberate less energy from their fuel. A nuclear explosion in space also radiates most of it’s energy as X-Rays which on earth you heat the atmosphere but in space mostly just do nothing. So their directly turning less nuclear energy into motion. Further, a smaller percentage of that force is then collected by the pusher plate compared to the Bell nozzle at the bottom of a traditional rocket. https://en.wikipedia.org/wiki/Bell_nozzle
Now as nuclear is so much more energy dense than chemical energy on paper they look better. But compare ISP’s and you see just how much energy is lost.
All of that can be simplified by just looking at ION thrusters ISP and vs the ISP of hypothetical Orion rockets. At that point it’s just a question of what percentage of mass is reaction mass vs the reactor + fuel and it turns out you can get reaction mass well over 80% making at least 60% more efficient before considering the pusher plate assembly’s mass vs physical ION engine’s mass.
PS: And that’s when looking at existing ION engine designs. Hypothetically, you could eject matter at a large fraction of C using energy giving orders of magnitude higher ISP, at the cost of lower energy efficiency etc.
It’s one of those designs that trades something considered largely meaningless, total engine weight, for something considered far more important exhaust velocity. This always ends up being a far worse trade off than simple models suggest.
PS: It could have very high acceleration before running out of fuel, but if the trip is going to take decades either way that’s not very important.