mlr3mbo
mlr3mbo picked its defaults from a benchmark study, not from taste
A side-by-side editorial comparison of bbotk and cmdstanr — release velocity, themes, recent moves, and the top alternatives to consider.
bbotk is generalizing from an optimizer toolkit into an evaluation framework.
bbotk is the black-box optimization backend behind mlr3 tuning: search spaces, terminators, archives, and an async layer built on rush. Recent releases pair steady async-API buildout with removal of the deprecated arguments that preceded it. The 1.9.0 release introduced EvalInstance as a base class for OptimInstance, separating evaluating an objective from optimizing one.
cmdstanr keeps adding fast approximations beside full HMC, and fighting Windows toolchains.
cmdstanr is the lightweight R interface to CmdStan, shelling out to the Stan binary rather than embedding it. The visible releases pair inference-method expansion, with laplace and pathfinder arriving in 0.7.0, against a continuous effort to make installation work on Windows. The most recent releases are dominated by CmdStan version compatibility and numerical fixes in the loo path.
bbotk is the black-box optimization backend behind mlr3 tuning: search spaces, terminators, archives, and an async layer built on rush. Recent releases pair steady async-API buildout with removal of the deprecated arguments that preceded it. The 1.9.0 release introduced EvalInstance as a base class for OptimInstance, separating evaluating an objective from optimizing one.
Two threads run through the visible history. The first is async optimization maturing: ArchiveAsync gained a full push/finish/fail vocabulary over rush tasks in 1.11.0, and 1.12.0 deleted the deprecated extra arguments it replaced. The second is dependency consolidation, with custom C hypervolume code handed to moocore and rush pinned to 1.0.0, trimming maintenance surface as the async path becomes the default.
The deprecation removals in 1.12.0 suggest the async archive API is now treated as settled; the next releases most likely build on EvalInstance rather than continuing to churn ArchiveAsync.
cmdstanr is the lightweight R interface to CmdStan, shelling out to the Stan binary rather than embedding it. The visible releases pair inference-method expansion, with laplace and pathfinder arriving in 0.7.0, against a continuous effort to make installation work on Windows. The most recent releases are dominated by CmdStan version compatibility and numerical fixes in the loo path.
The package tracks CmdStan closely, and its own work concentrates in two places. One is broadening the method surface so approximate inference sits beside sampling on the same object, extended in 0.8.0 by letting a completed fit supply initial values for the next run. The other is cutting installation friction, which reaches its conclusion in 0.9.0 with RTools45 supported and no additional toolchain setup needed on Windows. Dependency trimming, such as dropping RcppEigen for direct Eigen interop, runs alongside both.
The cadence is a CmdStan release followed by a compatibility release here, so expect the next to track a newer CmdStan and continue the effective-sample-size numerical work in the loo method.
Other Analytics products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either bbotk or cmdstanr.
mlr3mbo picked its defaults from a benchmark study, not from taste
loo keeps rewriting the diagnostics Bayesian modellers read off model comparison
mlr3fselect turned feature selection into an asynchronous, distributable job
lime survives on compatibility patches years after its research moment
mlr3measures is systematically retrofitting sample weights across every metric
mlr3cluster went from a handful of clusterers to covering the field
See all bbotk alternatives → · See all cmdstanr alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. bbotk is currently shipping more aggressively (velocity 2.5 vs 0.0), with 0 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. bbotk is currently shipping more aggressively (velocity 2.5 vs 0.0), with 0 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other Analytics products to evaluate alongside.
Top bbotk alternatives in Analytics are ranked by recent ship velocity. Browse the "bbotk alternatives" section above for the current picks, or visit /alternatives/bbotk for the full list with editorial commentary on each.
Top cmdstanr alternatives in Analytics are ranked by recent ship velocity. Browse the "cmdstanr alternatives" section above for the current picks, or visit /alternatives/cmdstanr for the full list with editorial commentary on each.