A few notes before we get to the eventual answer:
Resetting Portent won't affect much
All of the diviner's divination spells are quite bad to cast in combat, and if your medusa takes a turn off to do so (they each require at least an action,) that loss in action economy likely outweighs the usefulness of the portent. Acting optimally, this should never come up, so we can discount it in our calculations.
Without that ability, this is mostly just a spellcasting feature
Assume medusa only has one use of portent. This is almost certainly not enough to raise her challenge rating by a tier, so we can ignore it and focus on the spellcasting feature.
Spells and abilities affect challenge rating
This is the core of the question. The DMG suggests that spells affect challenge rating based upon how the affect the creature's offensive and defensive CR (DMG pg. 279.) Specifically, we need to focus on combinations of traits, since we know the CR of both of our base monsters.
Let's focus on defensive CR first. A medusa with mage armor has exactly the same AC as a normal medusa. No problem there. Stoneskin might be an issue if the medusa has the foresight to cast it before combat (they are a diviner, after all,) but since the medusa has fair odds of losing it on the first hit, it's probably not a big difference. Fly is a similar story. Spells probably don't affect your medusa's defensive CR much, since they don't interact very favorably with their inherent abilities.
Flipping the script, how does a diviner change if they gain the medusa's ability to turn foes to stone? If you assume that this ability is similar in effect to faerie dragon's superior invisibility (in that the players will usually not be able to see the medusa,) the DMG suggests we raise the effective AC of the diviner by 2 (DMG pg. 281.)
Looking at things from this perspective, though, we also need to apply the medusa's HP to the diviner, since we're now modifying that stat block instead of the medusa's. The diviner has quite low HP for its CR. Perhaps the designers were factoring in stoneskin after all! Honestly, I don't think we should, though. As appropriate as it would be for a medusa to have stony skin, I can't see it staying relevant in combat for very long. Since we're roughly doubling HP, much as stoneskin would, we're not changing anything from the designers' (assumed) calculations.
Offensive CR is a bit more interesting, since the medusa can restrain foes without requiring an action. Things get tricky here: do you expect your PCs to gaze upon the medusa? If so, then spells like fireball and ice storm become more damaging. If not, then this is no change from a diviner's normal offensive CR. Gaining the medusa's attacking features doesn't help the diviner much; they can already output as much damage per round with spells alone.
In summary, adding medusa abilities to a diviner increases their defenses (+2AC), and maybe their attacks (by an unknown and potentially very small amount).
Putting it together
We can see from the DMG that an effective AC increase of 2 is equivalent to increasing its defensive CR by 1 ("adjust the challenge rating suggested by its hit points up or down by 1 for every 2 points of difference.")
Since we got a total increase to defensive CR of 1, and not much from offensive CR, we can say a diviner with medusa abilities has its CR increased by either 0 or 1, for a total of 8 or 9.
Why is this so messy?
At this point, you're likely thinking that this is a lot of handwaving. And that's true. It's difficult to get exact numbers without making assumptions about uses of actions that will, in practice, turn out to be incorrect. What if the medusa goes unnoticed, and nukes the camp with a delayed blast fireball? What if they spend all their actions on trying to avoid getting hit, dealing no damage? What if the party only uses area of effect spells, negating the downside of not being able to see the medusa? What if they do look at the monster and die to just the passive ability?
Actual CR is dependent upon the abilities and strategies employed: everything else is estimates and guesswork. Or, to quote the DMG, "After seeing your monster in action, you might want to adjust the challenge rating up or down based on your experiences" (DMG pg. 275.)
So to your goals, that cause you to ask for a 'true' challenge rating:
- Appropriate treasure: This is a monster that is likely to be a recurring villain, given their ability to teleport away when things aren't going in their favor. I wouldn't recommend basing treasure off of CR in this case, but if you are, it really doesn't need to be exact.
- Experience rewards: This is similarly a tricky subject. If the medusa is routed, teleports away, and lives to fight another day, have the party "defeated" the monster? According to the DMG, yes (pg. 260.) Withholding experience here would likely feel worse than underestimating the amount when the medusa is eventually defeated. You're going to need to handwave this amount anyway: you're really looking at multiple sub-CR 10 encounters, if the medusa has a decent survival instinct.
- If your playgroup is happier when abilities are justified by CR, 9 is a reasonable estimate. So is 8, or 10 for that matter. Context matters more here than numbers, and how difficult the monster will be to defeat may have more to do with their allies and defenses.
Note: why isn't it CR 14?
Two answers here: Offensive CR and Defensive CR. Giving a monster twice as many abilities doesn't inherently enhance either of these. Defensive CR is based on effective AC and HP, which don't necessarily directly benefit from more abilities, and offensive CR is strictly limited by action economy. When you mash up two creatures this way, the best way to think about it is to assume the CR starts from the higher creature's rating; after all, it's usually going to use the best defensive and offensive abilities it has anyway, and those will often just be whatever the stronger monster had access to.
