Science Markets
Markets / IRON / Round 6

Round 6

One saved weekly assessment of Pure iron favours bcc over hcp at 330 GPa and 6000 K., kept as it was published. It is an estimate from three models reading the evidence, not proof the hypothesis holds, and it did not set any price.

The round
Published
Nov 16, 2026 09:11 AM EST
Estimate
20%
Range the models covered
7–27%
Confident binary
No unanimous verdict this round.
Methodology
panel-v9
Round hash
bf6bad28020b9aeb39ddb37bd585ef2153873df4490a2b875225bd1aad2a2b1c

The hash commits to this round's recorded bytes. It does not prove the wording is sound or that the science is right.

What each model said
OpenAI 0.25

openai/gpt-6.1-sol

The strongest retrieved calculations favor hcp, lowering my support for the proposed negative free-energy difference. However, the exact-state structure search loses the bcc basin, while quantitative free-energy results concern nearby pressures and retain electronic-structure uncertainty. Neither establishes the required independently reproduced, exact-state comparison with comprehensive 95% intervals; basin instability is explicitly nonterminal.

What would change this view: Independent constrained-basin calculations at exactly 330 GPa and 6000 K, with validated electronic-structure errors and wholly negative 95% intervals, would substantially raise my estimate.

Anthropic 0.07

anthropic/claude-sonnet-5.5

Available first-principles work (an ab initio TI study of bcc and hcp melting, and the HMA-DFT study of iron phases) finds hcp more stable than bcc near 323–360 GPa and 6000 K. The free-energy gap is only about 8–13 meV/atom, so it is small compared with method uncertainty. The criterion's NO condition (all eligible intervals wholly ≥0) is not clearly met by independent checked calculations, and bcc stabilization claims (Belonoshko) conflict with it. I therefore lean NO but it falls short of a rigorous settled result.

What would change this view: A rigorous, independently reproduced calculation with uncertainty below zero for G_bcc-G_hcp at 330 GPa and 6000 K would move me toward YES. Several independent calculations with intervals wholly above zero would firm up NO.

xAI 0.27

x-ai/grok-4.7

Recent ab initio thermodynamic integration finds hcp lower than bcc by about 8–13 meV/atom near 323–360 GPa, so G_bcc−G_hcp < 0 is not supported. That margin matches electronic-structure error, the state point is not exactly 330 GPa and 6000 K, and earlier studies disagree. No independent pair reports 95% intervals wholly below or above zero.

What would change this view: Two independent calculations at 330 GPa and 6000 K with 95% intervals for G_bcc−G_hcp wholly below zero, including electronic-structure error, would raise this.

The round note and its sources

Round 6: score 0.20

OpenAI estimated 0.25. The strongest retrieved calculations favor hcp, lowering my support for the proposed negative free-energy difference. However, the exact-state structure search loses the bcc basin, while quantitative free-energy results… Anthropic estimated 0.07. Available first-principles work (an ab initio TI study of bcc and hcp melting, and the HMA-DFT study of iron phases) finds hcp more stable than bcc near 323–360 GPa and 6000 K. The free-energy gap is only about 8–13… xAI estimated 0.27. Recent ab initio thermodynamic integration finds hcp lower than bcc by about 8–13 meV/atom near 323–360 GPa, so G_bcc−G_hcp < 0 is not supported. That margin matches electronic-structure error, the state point is not… The panel disagreed by 0.20, so read this average as soft. The three models did not return a unanimous confident binary, so the market is not decided.

Audit bundle

No audit bundle is recorded for this round. A bundle from another round is not substituted for it.

Back to IRON · How the current round reads