Opus 5 ties Sol, Terra matches Fable at a quarter the price
Sol, Terra, Opus 5 and Fable 5 on the same agent: a dead heat at the top, the same score at a quarter of the cost, and effort settings that can hurt.
Topic
benchmarks
Every article tagged benchmarks, newest first.
Kimi K3 is real, GLM-5.2 is a gamble, neither is a bargain
Kimi K3 ties GPT-5.6 on the independent coding index at 4.6 times the runtime. GLM-5.2's $1.40 tokens cost $6.51 per task. The Chinese model hype, measured.
Opus 5 is a great employee and a frustrating boss
Opus 5 matches Fable 5 on the benchmarks at half the token price, yet the first week of complaints is real too. Where it shines, and where it needs a boss.