theoretical-computer-science / Algebraic complexity, bilinear and tensor methods

The matrix multiplication exponent: is omega at most 9/4?

Theorem 1.1: for every ε>0\varepsilon>0 two complex n×nn\times n matrices can be multiplied with Oε(n9/4+ε)O_\varepsilon(n^{9/4+\varepsilon}) arithmetic operations, so ω≤9/4\omega\le9/4 over C\mathbb C. The companions prove α>0.465\alpha>0.465 over every field of characteristic zero, ω(1,0.709,1)<2.092\omega(1,0.709,1)<2.092, ω<2.258\omega<2.258 over all fields outside one uncomputed finite set of positive characteristics, and ω<2.371054886006746\omega<2.371054886006746 over every field. Not shown: anything about ω=2\omega=2, bit complexity, or a practical crossover size; the 9/4 proof gives no competitive finite matrix size and its transfer to positive characteristic is not claimed.

55Significance / 100
1Frontier events
0Verification tasks
0Recorded attempts

Temporal state

Current frontier

No reconciled state yet.

Append-only history

Frontier timeline

theoretical-computer-scienceOct 2, 2026Significance 55/100Registry: lean checked

The matrix multiplication exponent: is omega at most 9/4?

Prior state unknown→proved

Theorem 1.1: for every ε>0\varepsilon>0 two complex n×nn\times n matrices can be multiplied with Oε(n9/4+ε)O_\varepsilon(n^{9/4+\varepsilon}) arithmetic operations, so ω≤9/4\omega\le9/4 over C\mathbb C. The companions prove α>0.465\alpha>0.465 over every field of characteristic zero, ω(1,0.709,1)<2.092\omega(1,0.709,1)<2.092, ω<2.258\omega<2.258 over all fields outside one uncomputed finite set of positive characteristics, and ω<2.371054886006746\omega<2.371054886006746…

SourceReplayReproducedFormal proofStatement auditExternal checkExpert reviewPeer review

Research memory

Claims and attempts

Scoped claims

Source authenticated

Theorem 1.1: for every ε>0\varepsilon>0 two complex n×nn\times n matrices can be multiplied with Oε(n9/4+ε)O_\varepsilon(n^{9/4+\varepsilon}) arithmetic operations, so ω≤9/4\omega\le9/4 over C\mathbb C. The companions prove α>0.465\alpha>0.465 over every field of characteristic zero, ω(1,0.709,1)<2.092\omega(1,0.709,1)<2.092, ω<2.258\omega<2.258 over all fields outside one uncomputed finite set of positive characteristics, and ω<2.371054886006746\omega<2.371054886006746 over every field. Not shown: anything about ω=2\omega=2, bit complexity, or a practical crossover size; the 9/4 proof gives no competitive finite matrix size and its transfer to positive characteristic is not claimed.

Theorem 1.1: for every ε>0\varepsilon>0 two complex n×nn\times n matrices can be multiplied with Oε(n9/4+ε)O_\varepsilon(n^{9/4+\varepsilon}) arithmetic operations, so ω≤9/4\omega\le9/4 over C\mathbb C. The companions prove α>0.465\alpha>0.465 over every field of characteristic zero, ω(1,0.709,1)<2.092\omega(1,0.709,1)<2.092, ω<2.258\omega<2.258 over all fields outside one uncomputed finite set of positive characteristics, and ω<2.371054886006746\omega<2.371054886006746 over every field. Not shown: anything about ω=2\omega=2, bit complexity, or a practical crossover size; the 9/4 proof gives no competitive finite matrix size and its transfer to positive characteristic is not claimed.

Recorded attempts

Evidence graph

Connected research record