Which model leads Arena-Hard (GPT-4-1106 judge)?

Q: Which model leads Arena-Hard (GPT-4-1106 judge)?

DeepSeek-R1 tops the board at 92.3

Arena-Hard (GPT-4-1106 judge) Leaderboard — Model Scores

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

The field · 6 measured

53.692.3

6 models measured, most on the official card harness. DeepSeek-R1 tops the board at 92.3.

6 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 6Bars show 53.6–92.3 · ◆ solid = first-party or better · outlined = arm's-length

01DeepSeek-R1+5 altsDeepSeek92.3$2.0/M$4.0/M$0.033

02O1 Mini+4 altsOpenAI92———

03DeepSeek-V3+4 altsDeepSeek85.5$0.24/M$0.90/M$0.007

04Claude 3.5 Sonnet+4 altsAnthropic85.2$3.0/M$15.0/M$0.106

05GPT-4o+4 altsOpenAI80.4$5.0/M$15.0/M$0.124

06DeepSeek-R1-ZeroDeepSeek53.6———

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