Opus 5.5 beat Fable 5.1 on HAProxy port time and cost
Two models did the same C to Rust port, and the gap showed up in hours and dollars rather than pass rates.
Boris Cherny had Opus 5.5 and Fable 5.1 each port HAProxy from C to Rust. Both passed nearly all of HAProxy's tests. Opus 5.5 finished in 9.5 hours against Fable 5.1's 12, and did it for 51% less cost. Cherny also says Opus 5.5 has been his daily driver for the last few weeks, which is worth reading as the disclosure it is rather than as a neutral result.
Opus 5.5 finished in 9.5 hours compared to Fable 5.1's 12 hours, and for 51% less cost
This is the same head to head we covered earlier, now with the time and cost numbers attached. The interesting part is not that one model won. It is that the quality axis collapsed. Two different models each rewrote a large C proxy in Rust well enough to pass nearly all of its test suite, so the only thing left to compare was the meter.
The same pattern shows up on the Next.js leaderboard
Guillermo Rauch posted fresh Next.js evals with Opus 5.5, GPT-6 Sol and Fable 5.1 all landing at 97%, and Grok 4.7 at 94%. Rauch notes Grok is 2x to 7x cheaper. The Next.js account adds that among the three models tied at 97%, Opus ranks first on average cost.
So you get a three way tie at the top and a fourth model three points back for a fraction of the price. Picking between them is now a budget decision with a rounding error of capability attached, and the answer changes as prices and models move.
For working engineers
One port on one codebase is one data point, not a trend, and the Next.js numbers are a single benchmark on a single framework. But both point the same way. If the models you are choosing between all clear your test suite, the useful benchmark to run is your own, measured in wall clock time and spend on your repo, not a leaderboard position.
The volatility is the other half of the story. Theo, who has been running his own model comparisons all month, says the subscription economics have flipped in a few weeks.
Nobody in these sources has a theory for why the ordering keeps churning, and none of them claims the current standings will hold. The reasonable posture is to keep the switch cheap.
It's insane how much better the $200 Claude Code plan is compared to Codex right now.
Have the good opinion, not the loud one
One email a day. Under three minutes. What working engineers found out about coding with AI.