#benchmarking

All articles tagged with "benchmarking"

Rendering of a proposed apparatus for standardized testing of humanoid robot capabilities.

NIST proposes baseline benchmark for humanoid robots as investment race accelerates

NIST has proposed a standardized, low-footprint benchmark for humanoid locomotion and manipulation, aiming to create a common way to evaluate commercially relevant robot capabilities.

Key Takeaways

NIST says its proposal is the first standardized humanoid benchmark since the 2015 DARPA Robotics Challenge.
The benchmark focuses on low-footprint locomotion and manipulation tasks.

DT Editorial Team·May 30, 2026·via therobotreport.com

Cursor Unveils Composer 2.5 as a Lower-Cost Challenger in AI Coding

Cursor says its new Composer 2.5 coding model matches top benchmark results from larger rivals while operating at substantially lower token prices.

Key Takeaways

Cursor says Composer 2.5 was trained on 25 times more synthetic tasks than Composer 2.
The model reportedly scored 79.8 percent on SWE-Bench Multilingual and 63.2 percent on CursorBench v3.1.

DT Editorial Team·May 19, 2026·via the-decoder.com

GPT-5.5 Raises the Ceiling on AI Benchmarks but Keeps a Familiar Weakness

OpenAI’s GPT-5.5 has moved to the top of a major benchmark ranking and appears more token-efficient than its predecessor, but reporting cited in the source says the model still hallucinates at a high rate.

Key Takeaways

GPT-5.5 leads the Artificial Analysis Intelligence Index with 60 points.
Lower token consumption reduces the real-world price increase versus GPT-5.4 to about 20 percent, according to the source.

DT Editorial Team·Apr 25, 2026·via the-decoder.com

AI Soccer Betting Test Shows Frontier Models Still Struggle With Real-World Forecasting

A new benchmark built around Premier League betting found that every frontier model tested lost money over a season, underscoring the gap between strong performance on narrow tasks and messy real-world prediction.

DT Editorial Team·Apr 11, 2026·via arstechnica.com