Everything Model Zero does, in order

Solana Launch Dataset v15 · boosted decision trees · Model Zero training server

Same configuration as v0.18 (best so far on the headline benchmark), retrained on the newest dataset.

00:02:13 · trained on 8141 examples · $0 compute cost

Solana Launch Dataset v14 · boosted decision trees · Model Zero training server

Same configuration as v0.18 (best so far on the headline benchmark), retrained on the newest dataset.

00:02:20 · trained on 1808 examples · $0 compute cost

Solana Launch Dataset v13 · boosted decision trees · Model Zero training server

Adds global fees paid : total fees and Jito tips, fees per minute, fees relative to volume, transactions per paying wallet and failed-transaction share — then lets automatic discovery choose which change-signals to add on top.

00:03:04 · trained on 6421 examples · $0 compute cost

Solana Launch Dataset v12 · boosted decision trees · Model Zero training server

Automatic signal discovery: Model Zero generated candidate signals (how every measure changed since the previous snapshot), tested each on held-out folds of the training launches, and kept only the ones that improved its predictions.

00:03:24 · trained on 6411 examples · $0 compute cost

Solana Launch Dataset v11 · boosted decision trees · Model Zero training server

Same signals with deeper trees, to learn interactions such as 'bundle still holding while snipers exit'.

00:01:39 · trained on 6361 examples · $0 compute cost

Solana Launch Dataset v10 · boosted decision trees · Model Zero training server

Adds launch structure — bundled buys in the creation slot, snipers in the first 10 seconds and how much they've sold, first-minute buying, organic-volume checks (volume per wallet, back-and-forth wallets, tiny bump trades), KOL buys, and the dev's launches across all pump.fun coins.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:04 · trained on 3502 examples · $0 compute cost

Solana Launch Dataset v9 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:13 · trained on 3332 examples · $0 compute cost

Solana Launch Dataset v8 · boosted decision trees · Model Zero training server

Same configuration as v0.05 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:11 · trained on 2752 examples · $0 compute cost

Solana Launch Dataset v7 · boosted decision trees · Model Zero training server

Same configuration as v0.05 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:03 · trained on 2187 examples · $0 compute cost

Solana Launch Dataset v6 · boosted decision trees · Model Zero training server

Same configuration as v0.04 (best so far on the headline benchmark), retrained on the newest dataset.

00:00:03 · trained on 679 examples · $0 compute cost

Solana Launch Dataset v5 · boosted decision trees · Model Zero training server

Adds the realized track record (inside Model Zero's sample, before the snapshot) of the wallets buying the launch — how many are known, profitable, losing or high-frequency.

00:00:04 · trained on 1025 examples · $0 compute cost

Solana Launch Dataset v4 · boosted decision trees · Model Zero training server

Deeper trees (depth 3), more of them (300) with a lower learning rate.

00:00:01 · trained on 594 examples · $0 compute cost

Solana Launch Dataset v3 · boosted decision trees · Model Zero training server

Switches to gradient-boosted trees (depth 2, 120 trees), which can learn thresholds and interactions a linear model cannot.

00:00:00 · trained on 110 examples · $0 compute cost

Solana Launch Dataset v2 · linear model · Model Zero training server

Adds trade-flow, holder-estimate, creator-history and social features (all 36 features in set f1).

00:00:00 · trained on 60 examples · $0 compute cost

Solana Launch Dataset v1 · linear model · Model Zero training server

First learned model: logistic regression on price and volume only (12 features).

00:00:00 · trained on 60 examples · $0 compute cost

Solana Launch Dataset v1 · baseline (averages) · Model Zero training server

Baseline. Predicts each outcome's historical base rate and ignores the token entirely. Every later version has to beat this to be worth anything.