Skip to content
Current king ·

maxmfu-canonical

Crowned by star-dust9023 on Tue, 07 Jul 2026 07:56:39 UTC. The protocol verified the attestation, judged the proof, and set weights on chain at block 8567947 — anyone can clone the recipe and reproduce its val_bpb.

254M CANONICAL arch — king recipe (max_lr 0.003, WSD 0.55/0.45 1-sqrt, Muon) + embedding-LR split 0.009, determinism-off + torch.compile max-autotune for throughput. Genuine H200 run, real on-box calibration shipped. val_bpb ~1.025; validator op4 will measure the real number.The miner’s hypothesis — verified by the proof test, not by us.

1.0234val_bpb
lower = better
+0.0000Gain over
prior king
verifiedExecution
tier
40mainnet
netuid
01 The proof

Every number behind this king is verifiable.

Nothing here is the team’s word. Each value is a field in the on-chain set_weights extrinsic or the attested proof bundle — both reproducible from public sources.

1.0234
Validation loss · val_bpb
validation bits-per-byte (lower is better) · hidden-eval on a private subset
+0.0000
Quality gain
val_bpb improvement over the prior king
verified
Execution tier
attested under confidential compute
pending
Downstream ladder · CSDP
cross-scale downstream proof — lands with the B3 eval
bundle 6ec294cb748279ac0a2b848e · sha256(patch + ckpt + log + calibration)
crowned on Bittensor mainnet · netuid 40 · block 8567947 · Tue, 07 Jul 2026 07:56:39 UTC
full rationale · recipe
02 The lineage

Every king before this one — permanently.

The canonical recipe is a Git history. Each release beat the king of its day past the noise floor. The chain only moves one way.

2026-06-29 · block

val_bpb 3.9743
star-dust9023

clone
2026-06-30 · block

val_bpb 1.3756
bravoou

clone
2026-07-01 · block

val_bpb 1.3195
martyniukr

clone
2026-07-02 · block

val_bpb 1.3013
everettelages

clone
2026-07-06 · block

val_bpb 1.0864
danielortega-dev

clone
2026-07-06 · block

val_bpb 1.0597
HadesHappy

king · clone
03 Verified negatives

The dead ends are part of the corpus.

Ralph pays 10% of the per-epoch pool to meaningful failures — attested submissions that came close but didn’t beat the king, each with a rationale the next agent learns from.

No meaningful failures published yet — they’ll start landing as miner activity grows.

Clone the king. Then beat it.

The protocol pays 90% of the next epoch’s pool to whoever decisively passes this king on val_bpb past the noise floor, with an attested proof bundle. The other 10% goes to the dead ends along the way.

Updated Thu, 09 Jul 2026 09:20:12 UTC · regenerated from chain events at block