Energetic nuclear reactions, such as fission (of fissile materials) or fusion (of light elements) or annihilation (of anything) doesn't cause much of an explosion by itself, because there's simply not enough "stuff". In the extreme, complete annihation of matter can't be an explosion at all... just a very bright, very brief gamma-ray source.
Instead, what happens is that the intense x-ray or gamma radiation, combined with energetic fast particles such as neutrons, alpha particles and heavy fission fragments, are eventually absorbed by regular matter around them which heats up very rapidly and expands violently, forming the actual explosion. The characteristic fireball of a nuclear explosion is superheated and partially ionised air and dust, for the most part.
If your bullet contains a suitable confinement mechanism for a microgram of antimatter, the total useful amount of energy released is about 90MJ, of which about 60MJ is released in the form of highly penetrating 200MeV gamma rays. Given the stuff humans are made of, they have a mass attenuation coefficient of about 0.01-0.02cm2/g against photons of that wavelength, which means that over a 10cm distance only 18% of the energy will be absorbed.
Now, what follows is only slightly better than a wild guess... bulk gamma ray absorbtion is hard to calculate (because integration over a sphere-intersecting-with-a-human is awkward and because meat plasma is less absorbing than meat), but here goes anyway. I'll work it out for the most penetrating gamma rays, meaning that the actual amount of absorption will be higher which should make up for ignoring the plasma phase transition.
Anyway. If you shoot someone right in the middle of their torso, and the bullet detonated half way through, a sphere of meat 10cm in radius will absorb at least 16MJ of energy, giving an energy density of about
390MJ/kg, about a hundred times more than the equivalent mass of TNT, so it'll definately expand energetically as it turns to superheated and partially ionised vapour, aka a nuclear fireball. The yield over that small spherical volume is equivalent to about 4kg of TNT. In reality of course the explosion could be more energetic... a lot more energy will be absorbed from gamma rays travelling up and down the long axis of the body, and yet more will be absorbed by the target's clothing and equipment and immediate surroundings. Note that anything nearby that survived the blast will receive a nasty dose of radiation... I'm not going to calculate how nasty (because it is hard) but it will certainly be bad news.
Here's a video of 8kg TNT equivalent of high explosive. It probably won't be dissimilar to the effect of the antimatter bullet... if anything, the bullet I've look at would be much more destructive (max equivalent yield perhaps 21.5kg TNT) but this should give you a good idea of how things will go.
The explosive was detonated inside a vehicle the size of a large family car or small van, to give you some idea of scale. The dark square object in the bottom right of the fireball was one of the doors. I suspect the actual fireball would start out brighter (perhaps blindingly white, then bright yellow-white) and then be less orange afterwards, given the elemental composition of meat, but I couldn't say for sure.
If you do make antimatter bullets, it would be inadvisable use them at close range.
