Artificial Analysis billboards rank leading models by intelligence and cost per task
ArtificialAnlys · x · 2026-07-24
Artificial Analysis says model cost efficiency is becoming a primary performance dimension.
- Their new San Francisco billboards show an Intelligence Index vs. Cost per Task chart for leading models.
- Fable 5 tops the index at 60 but averages $2.75 per task.
- Grok 4.5 (high) scores 54 at $0.31 per task, about 9× cheaper, and still sits on the Pareto frontier.
- DeepSeek v4 Flash scores 44 but costs $0.04 per task, about 69× cheaper than Fable.
The chart’s point: on the Pareto frontier, no model is better on both intelligence and cost, so models outside it are paying more for less.
Related event: Artificial Analysis Puts AI Cost-Performance on SF Billboards(2 posts)→
More from Models
- Mathematicians debate AI math writing: abysmal today, improving shockingly fast — miniapeur · 2026-10-27
- LessWrong: GLM models show outlier chain-of-thought controllability vs Kimi and DeepSeek — tomekkorbak · 2026-10-27
- AI deference trap: agent spots medical red flag but stays silent to avoid questioning the clinician — echen · 2026-10-27
- The trust paradox: only the least trustworthy LLMs will do the sketchiest things — AaronBergman18 · 2026-10-27
- Claude's refusal to type one-time passwords is a real security impediment — AaronBergman18 · 2026-10-27
- repligate argues context full of noise is genuinely 'hard' for LLMs, mulls an info-theory model of breaks — repligate · 2026-10-27