UFC Belgrade Preview: Medic vs Rodriguez Predictions, Full Card Breakdown and Every AI Pick

The UFC has never been to Serbia before, and it's making up for lost time with a stacked Saturday: UFC Fight Night: Medic vs. Rodriguez, August 1 at Belgrade Arena. Set your alarm accordingly if you're stateside, because this is an afternoon card. Prelims start at 10 AM ET and the main card goes at 1 PM ET, which means you can watch a full night of Serbian fights and still have your Saturday evening.

Every fight on this card has been through our full pipeline: all twelve AI models on the leaderboard forecast every bout line-blind, and the field currently sits 88.9% on unanimous calls (24-3) since we started grading UFC in public. Below is the main card fight by fight, what the models see, and the two spots where the field and the betting market flatly disagree.

Uros Medic vs Daniel Rodriguez, welterweight main event

The story writes itself: Medic is the first Serbian to headline a UFC card, doing it in Belgrade, riding three consecutive first round knockouts. The 33-year-old has a 100 percent finishing rate across 13 wins, and his stat line reads like a man in a hurry: 5.59 significant strikes landed per minute at a 60 percent clip, which is elite accuracy for someone throwing that much.

Rodriguez is the test the UFC likes to give hometown headliners. The 39-year-old Californian has seen everything across 25 fights, he's also on a three-fight win streak, and his style is the exact opposite of a sprint: 7.20 strikes landed per minute but 5.40 absorbed, a volume brawler who wants the fight long and messy. The fight breaks down to one question, and it's the question of the whole card: what happens when Medic's first-round blitz meets a man who has made a career out of surviving it. This is also Medic's first five-round fight, and Rodriguez's pace is precisely the thing that punishes a fast starter in rounds four and five.

Our field is decided but not comfortable. All twelve models take Medic, and all twelve call it by KO/TKO, but their average win probability is 65.6% against a betting market that prices him at 76%. Read that carefully: the AIs agree with the market's direction and think the price is too steep. When every model lands below the market number, the value, if there is any, lives on the underdog's side of the line.

Jan Blachowicz vs Navajo Stirling, light heavyweight

This one is the field's loudest opinion on the card. Blachowicz is a former light heavyweight champion, a legend at 43, and he has not won a fight in more than four years: two losses and two draws since 2021. Stirling is a 10-0 replacement opponent, twelve years younger, with a stat line that reads like a wave arriving: 6.24 strikes per minute landed against only 2.62 absorbed, with a 79-inch reach.

There's a scheduling wrinkle that matters here too. Stirling is a replacement opponent, stepping in during fight week preparations that Blachowicz originally aimed at a different matchup, and late opponent changes historically hurt the older, more game-plan-dependent fighter far more than the young athlete who just fights. Blachowicz's entire path to victory runs through his famous legendary Polish power and a veteran's clinch game, and the numbers say the window for landing it is narrow: he lands 3.55 strikes per minute to Stirling's 6.24, and gives up two inches of youthful movement everywhere else.

All twelve models pick Stirling. Not eleven, twelve, and ten of them call it by KO/TKO. The market agrees at 73%, so there's no pricing argument here, just a stark generational read. The models have been consistently unsentimental about aging champions all season, and this is the cleanest version of that test we've had yet: a name everyone in the arena knows against a prospect the algorithms love. We've written before about how model consensus usually just marks the obvious favorite, and that applies here, but the method unanimity is the part with teeth. Ten KO calls against a former champion with that much championship-round durability behind him is a bold, checkable claim. If Blachowicz wins, it's the biggest miss our unanimous-call record has taken, and it will be graded in public Sunday morning like everything else.

Aleksandar Rakic vs Marcin Tybura, heavyweight

Here's the fight where our field and the betting market genuinely split, and it's the one we'd watch most closely if you care about whether AI models can out-read a price. Rakic, an Austrian with Serbian roots who will hear the loudest cheers of the prelim-to-main stretch, is moving up to heavyweight after 20 fights at light heavyweight. Tybura is the definition of an established heavyweight: 38 professional fights, 40 years old, and steadfast in a grinding, clinch-heavy style backed by a 1.30 takedown average.

The market is convinced the athleticism travels: Rakic is priced at 75%. Our models are not convinced at all. Only five of twelve take Rakic, and the field's average probability on him is 46.2%, which functionally makes Tybura the field's lean at a +300 shape price. Seven models see a 34-year-old in his first fight at a new weight, carrying a long injury-interrupted stretch at light heavyweight, against a man who has never been athletic and never needed to be. Divisional debuts against entrenched veterans are exactly where markets overrate the shinier name. Someone is wrong by nearly thirty points of win probability here, and it's the widest field-versus-market gap we've graded on any UFC fight this summer.

Dusko Todorovic vs Robert Valentin, middleweight

The second Serbian storyline of the main card, and the second spot where the field is fading the price. Todorovic gets his homecoming at 32, with real punching volume (4.73 landed per minute at 57 percent accuracy) and real defensive holes (4.40 absorbed, 42 percent takedown defense). Valentin is coming off his first UFC win, an emotional first-round submission in April, and his profile is the inverse: low output at 1.30 strikes per minute, but a 1.80 submission average that turns Todorovic's takedown defense into the whole fight.

Eleven of twelve models take Todorovic anyway, and here's the interesting part: the market makes him the underdog at 41%. That's the near-unanimous-underdog shape our strategy board specifically hunts, the rare alignment where the field backs a fighter the market is against, and on the MLB side that exact shape has been one of the most profitable cuts we track. The models are betting that home-crowd Todorovic keeps this standing, where his volume advantage is enormous and Valentin's 1.30 strikes per minute cannot keep pace on the scorecards. The market is betting that Valentin only needs one takedown against a 42 percent defensive wrestler to find the submission he found in April.

The method votes tell you which way our field leans on that argument: eleven KO/TKO calls, one submission, zero decisions. Nobody in the field thinks this goes long, which matches two fighters whose last eight combined fights have almost all found a finish. Serbian crowd, live underdog price, and a genuine grappler-versus-striker fault line make this the main card's best pure watch even before the model angle. If you only track one non-headline result against our board Sunday, make it this one.

Vlasto Cepo vs Gilbert Urbina, middleweight

Cepo is the purest hometown story on the card: a 31-year-old Serbian veteran, 14-3 with a 100 percent finishing rate, getting his UFC debut in his own country. He has no octagon stat line yet, which makes him exactly the kind of fighter AI models have to price on record and context alone. Urbina, the TUF alum, has the numbers but they cut both ways: decent output, 4.62 absorbed per minute, and a takedown defense that has not stopped a single attempt in his UFC run.

Eleven of twelve models back Cepo and all twelve method calls say KO/TKO, in line with a market that has him at 74%. Here's the honest caveat we'd attach to that confidence: UFC debuts are the models' blind spot in both directions, because the packet they read has no octagon-level data for the debutant, and regional records can flatter anyone. The field is essentially pricing Cepo on a 14-3 record built entirely on finishes, a two-fight winning streak, and the shape of his opponent, and the shape of his opponent is the argument. Urbina absorbs 4.62 significant strikes a minute and has yet to stop a takedown in the UFC, which means every path in this fight runs downhill for the debutant. When a lifetime finisher meets a man with those defensive numbers in front of twenty thousand of the finisher's countrymen, the chalk logic is not complicated. It's just less certain than a 74% number makes it look, and we'll know by mid-afternoon Saturday whether the models priced their own blind spot honestly.

The prelims worth your morning coffee

Three picks stand out from the 10 AM ET portion. The field is perfect-twelve on Mateusz Rebecki against Kyle Prepolec at lightweight, and it's the most confident call on the entire card at an 81.7% average, powered by Rebecki's 3.05 takedown average against a fighter with 53 percent takedown defense. All twelve also take Ludovit Klein over Tofiq Musayev, every single one by decision, which is the models telling you Klein's 76 percent takedown defense neutralizes the 36-year-old Musayev's wrestling-first approach for fifteen minutes. And keep an eye on Michael Oliveira against Oban Elliott: eleven of twelve back the undefeated 9-0 Brazilian and his absurd 8.40 strikes per minute, the same lopsided-prospect shape as the Stirling call two fights later.

How to follow it with us

Every pick above is locked, timestamped, and will be graded Sunday morning, wins and losses alike, on the UFC board and the leaderboard. The unanimous calls feed The Unanimous Cage Call on our Pro board, which enters the weekend 24-3. And if you want to see how any individual model has done head-to-head against another on fights like these, that lives on the versus pages.

First prelim 10 AM ET, main card 1 PM ET, Saturday August 1. See you Sunday for the grades.