Which model leads MLE-Bench?

Q: Which model leads MLE-Bench?

Claude Sonnet 5 tops the board at 66.9

MLE-Bench Leaderboard — Model Scores

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

The field · 7 measured

39.266.9

7 models measured, most on the official card harness. Claude Sonnet 5 tops the board at 66.9.

7 measured·Lifecycle Updated Oct 8, 2026Score basis

#◆Vendor$/M in$/M out

Leader — ignites whiteField — fill cools with scoreIndependently verified only0 of 7Bars show 39.2–66.9 · ◆ solid = first-party or better · outlined = arm's-length

01Claude Sonnet 5Anthropic66.9$2.0/M$10.0/M$0.090

02Gemini 3.6 Flash+2 altsGoogle63.9$0.75/M$3.8/M$0.035

03Gemini 3.5 FlashGoogle49.7$1.5/M$9.0/M$0.106

04GPT-5.6 LunaOpenAI47.6$0.20/M$1.2/M$0.015

05Grok4.5SpaceXAI43.2$2.0/M$6.0/M$0.093

06Gemini 3.1 ProGoogle42.6$2.0/M$12.0/M$0.164

07Gemini 3.5 Flash Lite+1 altGoogle39.2$0.30/M$2.5/M$0.036

◆ 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: 7 models scored · 0 independently verified · 3 vendor cross-reference · 4 vendor-reported · 0 with source disagreement. How these tiers are assigned