Which model leads SWE-bench Verified (hard subset)?

Q: Which model leads SWE-bench Verified (hard subset)?

Claude Sonnet 4.5 tops the board at 45.3

SWE-bench Verified (hard subset) Leaderboard

Pro Mode — raw scores, one benchmark under the lensofficial card harness

The field · 3 measured

39.545.3

3 models measured, most on the official card harness. Claude Sonnet 4.5 tops the board at 45.3.

3 measured·Lifecycle Updated Oct 7, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 3Bars show 39.5–45.3 · ◆ solid = first-party or better · outlined = arm's-length

01Claude Sonnet 4.5Anthropic45.3$3.0/M$15.0/M$0.199

02Claude Opus 4.1Anthropic43.8$15.0/M$75.0/M$1.027

03Claude Opus 4Anthropic39.5$15.0/M$75.0/M$1.139

◆ solid = independently verified · aggregator attested · vendor attributed — outlined = cross-reference · source-attributed. Dimmed ◆ = confidence inferred. “+N alts” = alternate disclosures in the drawer with why-demoted. Prices per million tokens; $/point = input price ÷ score. Every row opens its model dossier.

Verification: 3 models scored · 0 independently verified · 3 vendor-reported · 0 with source disagreement. How these tiers are assigned