Theo says a Fable 5.2 game build cost $42 and Opus 5.5 did better
One run each, judged on how the two games felt to play rather than on a scored benchmark.
Theo ran Fable 5.2 on a 3D game build and reports it finished in 76 minutes for about $42 in tokens. He calls it one of the best runs he has seen on the project, which he refers to as Fishslop, and then immediately qualifies it. In his words the result was not significantly better than Opus 5.5, especially for the price.
One of the best Fishslop runs I've seen for sure, but not significantly better than Opus 5.5 (esp. for the price)
What he compared
The comparison is to an earlier Opus 5.5 build of the same thing. Theo gives Fable 5.2 the graphical edge on one specific count, it did an "open tank" thing the Opus version skipped. Everything else goes the other way. He says the Opus build feels more playable, that it got the gameplay loop better, and that controls and performance are better too. His conclusion is not that one model wins outright but that the two would work well together on 3D game type tasks.
In a follow-up he was blunter about the failure mode, saying Fable 5.2 was making mistakes he would not have expected from Opus.
How much weight this carries
This is one run per model on one project, scored by the person who commissioned both and judged on how the games felt in his hands. There is no eval, no rubric, no repeat trials, and the two runs did not happen under matched conditions that Theo describes. The $42 figure is the only hard number in the comparison, and it is attached to the Fable 5.2 run only. He does not post what the Opus 5.5 run cost, so the price complaint is directional rather than a measured ratio.
For people building on it, treat this as a taste verdict with a receipt attached rather than a head-to-head. The useful part is the shape of the failure, a build that looked better in stills and played worse, which is exactly the thing a static diff or a screenshot-based eval will not catch. If you are routing game or simulation work by model, the thing to measure is the loop someone actually plays, not the frame they screenshot.
Fable 5.2 is making a lot of stupid mistakes I would not have expected from Opus tbh
Know what Jev is good at before you bet on it
One email a day. Under three minutes. What builders found out about Jev yesterday.
