Which model leads ResearchRubrics?

ResearchRubrics Leaderboard — Model Scores

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

The field · 12 measured

54.376.2

12 models measured, most on the official card harness. Kimi K3 tops the board at 76.2.

12 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 12Bars show 54.3–76.2 · ◆ solid = first-party or better · outlined = arm's-length

01Kimi K3Moonshot76.23P$0.58/M$12.3/M$0.085

02GPT-5.6 SolOpenAI73.83P$4.0/M$20.0/M$0.163

03Claude Opus 4.8Anthropic73.53P$5.0/M$25.0/M$0.204

04GLM-5.2Z.ai71.13P$1.4/M$4.4/M$0.041

05Step 3.5 Flash+2 altsStepFun65.3$0.10/M$0.30/M$0.003

06GPT-5.5OpenAI643P$5.0/M$30.0/M$0.273

07GLM-4.7+2 altsZ.ai62$0.60/M$2.2/M$0.023

08MiniMax M2.1+2 altsMiniMax60.2$0.30/M$1.2/M$0.012

09Kimi K2.5+2 altsMoonshot59.5$0.45/M$2.3/M$0.023

10Kimi K2 Thinking+2 altsMoonshot56.2$0.60/M$2.5/M$0.028

11DeepSeek-V3.2+2 altsDeepSeek55.8$0.28/M$0.42/M$0.006

12MiMo V2 Flash+2 altsXiaomi54.3$0.10/M$0.30/M$0.004

◆ 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: 12 models scored · 0 independently verified · 10 vendor cross-reference · 2 vendor-reported · 0 with source disagreement. How these tiers are assigned