GateTruth

RTL model leaderboard

Signed PPA-aware evaluation results.

Track A (generation)

#ModelSiliconScorePass@1Mean PPAPassedCostTokensDateRun
1gpt-5openai / default48.8371.7%1.0236x43/60$2.318627320,0192026-07-30OFFICIAL
2claude-opus-4-8anthropic / default46.9970.0%1.0069x42/60$1.600930186,7462026-07-30OFFICIAL
3claude-sonnet-4-6anthropic / default42.0963.3%0.9968x38/60$0.856386148,1782026-07-30OFFICIAL
4gpt-5-miniopenai / default35.2353.3%0.9910x32/60$0.351382263,8472026-07-30OFFICIAL
5claude-haiku-4-5-20251001anthropic / default33.7850.0%1.0152x30/60$0.292247147,0192026-07-30OFFICIAL
6google/gemini-2.5-proopenrouter / default31.0646.7%0.9984x28/60$3.434077420,8052026-07-30OFFICIAL
7meta-llama/llama-4-maverickopenrouter / default24.6536.7%1.0084x22/60$0.027048118,6072026-07-30OFFICIAL