Slightly improved zero-free half-planes for the quasi-Riemann hypothesis
The quasi-Riemann hypothesis asks for a fixed B<1 such that zeta and L-functions have no zeros with Re s > B. A recent argument attributed to OpenAI gave a uniform 7/8 zero-free half-plane for finite-order Hecke L-functions over Q(sqrt(-3)), and transferred it to Dirichlet L-functions, using the fixed parameters b=1/8 and ell=1/6.
The same cubic-theta averaging strategy is followed, but the averaging parameters b and ell are varied. The direct estimate and the character-count estimates are tracked as functions of these parameters. The resulting exponent comparison is then optimized. The zero-free conclusions are formalized in Lean 4.
A small perturbation of ell gives B_r = 34999/40000. Optimizing both parameters gives B_new = (1507 - 2*sqrt(921))/1653 ≈ 0.874957, which improves 7/8 by about 4.29e-5. Both bounds are formalized in Lean for the Hecke family and for its Dirichlet and zeta consequences.
| Bound | b | ell | Value |
|---|---|---|---|
| Original | 1/8 | 1/6 | 0.875 |
| B_r | 1/8 | 1/6 + 1/10000 | 0.874975 |
| B_new | -4/29 + 230√921/26709 | (33+8√921)/1653 | 0.874957069799... |
