Which model leads AceBench?

Q: Which model leads AceBench?

Kimi K2 Instruct tops the board at 76.5

AceBench Leaderboard — Model Scores

Pro Mode — raw scores, one benchmark under the lensthird-party eval harness

The field · 5 measured

70.576.5

5 models measured, most on the third-party eval harness. Kimi K2 Instruct tops the board at 76.5.

5 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 5Bars show 70.5–76.5 · ◆ solid = first-party or better · outlined = arm's-length

01Kimi K2 Instruct+3 altsMoonshot76.5VENDOR$0.57/M$2.3/M$0.019

02Claude Sonnet 4+1 altAnthropic76.2$3.0/M$15.0/M$0.118

03Claude Opus 4+1 altAnthropic75.6$15.0/M$75.0/M$0.595

04DeepSeek-V3+1 altDeepSeek72.7$0.24/M$0.90/M$0.008

05Qwen3 235B A22B+1 altAlibaba70.5$0.70/M$2.8/M$0.025

◆ 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: 5 models scored · 0 independently verified · 4 vendor cross-reference · 1 vendor-reported · 0 with source disagreement. How these tiers are assigned