https://www.canonrumors.com/this-is-just-a-sunday-reminder-of-the-canon-eos-r5-specifications/https://www.canonrumors.com/canon-eos-r5-launch-price-will-be-below-4000-usd-cr3/This is looking very good.
The elephant in the room is its AF performance. 45MP and 20fps is all well and good, but the 20fps part of it is totally meaningless unless the AF can keep up with the action. You don't get a camera to shoot landscapes at 20fps. Is this an A7r competitor, or an A9r?
If it has A7-level AF, matching the A7r4 (which will be a year old by the time the R5 is available), I'd give the edge to Sony, if only due to (current) better lens selection and the fact that the A7r5 is likely to come much sooner than whatever succeeds the R5. This would still represent a big leap in performance for Canon mirrorless cameras (granted, the previous models have all been entry-level bodies, not 5D or 1D replacements), and the new competition would likely spur Sony into a new round of innovation and performance improvement, like what we saw in the early days of Sony full-frame mirrorless, where bodies came thick and fast and performance improved in leaps and bounds. Maybe not for the A7r5 (which is likely already mostly developed, although you can do a lot with firmware), but certainly for the A7r6, or a possible A9r.
If it has A9-level AF, it becomes much more interesting. This would put Canon in the lead with a unique camera, with a deficit in lenses that will grow smaller over the next few years. In one body, it would combine the capabilities of action cameras with those of high-resolution bodies - just as Canon had done, to an extent, back in the days of the 1Ds3 - and provide sports and wildlife photographers with the ability to heavily crop action shots in the way that no other camera can, either due to lack of resolution (the various 20-24MP action bodies) or more limited AF capability (the various high-resolution bodies). This would leave Sony playing catch-up - no doubt they'd be forced to release an A9r, but they'd be starting from behind, although Canon would still have to release some supertele primes and zooms for their mirrorless mount in order to compete with Sony in this area. But one effect would be that it would likely put people on notice and stall investment in the Sony system, with potential buyers choosing to delay any new, non-vital purchases (especially in high-value gear that will be used for a long time, such as supertele primes) until it became clearer which side (if any) would take a meaningful lead in action photography.
I doubt sensor performance will be a major issue. Canon has made a lot of progress over the last few years - their current sensors are respectable, if not class leading at base ISO. And, for action photography in the typical 400-12800 ISO range, there's no real difference in output quality between Canon sensors and Sony Exmor sensors - it's only at base ISO that Canon sensors have tended to fall behind? There's no reason to suspect that the sensor Canon puts into this camera will be so bad it would cause people who would otherwise consider the R5 on its other merits to stay away, or so good it leaves Exmor in the dust. Basically, sensor performance is unlikely to be a distinguishing factor between the R5 and its competitors - the battle will be fought on other grounds. (Does anyone actually care about the ISO 50k+ range anyway? At that end, it's just unusable (Canon) vs unusable (Sony) vs unusable (Nikon) anyway.)