TokenPost topic
Benchmark
6 TokenPost stories are tagged Benchmark. The newest appear first.
Latest Benchmark stories
- Mind2Act World Benchmark Finds 12% Success on Combined Robotics Tasks
The benchmark spans 180 simulated configurations. Its strongest agent scored higher when reasoning and execution were tested separately.
· Technology
- Stripe Benchmark Finds Agents Satisfy Every Payment Requirement in 29% of Runs
The open-source evaluation covers 11 reproducible environments, including browser-checked full-stack tasks. Agents averaged about 78% of the required criteria.
· Technology
- FrontierCS 2 Finds No AI Agent Passed 16 Research-Coding Tasks
The 48-task benchmark spans 13 computer science domains. Claude Opus 5.5 led the initial evaluation with a 58% pass rate.
· Technology
- Artificial Analysis Finds 770-Point Elo Gap Across 40 AI Models
GPT-4o (Nov. ’24) scored 1,467 Elo on sycophancy, while Grok 4.7 (Xhigh) scored 697. The benchmark tests 30 personality traits with 900 questions.
· Technology
- Benchmark Says AI Growth Is Straining Venture Firms' Exclusivity Rules
Partner Chetan Puttagunta expects startup overlap to increase as software becomes cheaper to build. Benchmark says exclusivity matters most at a company’s earliest stage.
· Business
- Hone Launches AI-Agent Platform With $60 Million Seed Round
The San Francisco startup is developing persistent AI systems that pursue business goals over weeks or months instead of completing isolated tasks.
· Business