Hacker News · AI·· 14 h agoAI score38
Vals AI Benchmark Comparison Ranks Claude, Gemini and GPT Models Across Legal, Finance and Coding Tasks
Vals AI – Compare AI models across real-world benchmarks
AI summary
Vals AI published a cross-model comparison on its Vals Index, a GDP-weighted benchmark spanning finance, coding, legal and tax, with top scores clustered at 70.14%–71.31%. The five compared models, from Anthropic and Google, are priced between $2/$10 and $8/$40 per 1M input/output tokens, all with 1M context windows. Per-task cost on the Vals Index ranges from $7.405 to $24.77.
Source:Hacker News · AI · vals.ai