Artificial Analysis updates Coding Agent Index with reward hacking corrections

Artificial Analysis updates its Coding Agent Index with Terminal-Bench v2.1, zeroing out reward hacking attempts and fixing 28 of 89 benchmark

Raccontata dacryptobriefing.com