GDPVal
GDPVal is a benchmark designed to evaluate the comprehensive capabilities of AI language models, covering performance assessment across a variety of real-world task scenarios.
Timeline (last 90 days)
The new model achieved 1,620 points on GDPVal (compared to 735 for the previous generation), improved on OSWorld from 15.7% to 72.4%, on Terminal Bench from 0.0% to 39.2%, and on Humanity's Last Exam from 10.2% to 45.9% without tools and from 18.7% to 57.4% with tools.
The Tongyi Qianwen Flash edition starts at $0.03 and $0.13, and GLM scored 1,647 points on GDPVal.
Anthropic states that Haiku 5.5 outperforms GPT-6 Luna on certain metrics including GDPVal and Agentic Coding.
On the GDPVal benchmark, GPT-6 Luna scored 1437, lower than Haiku 5.5's score of 1620.
On the GDPVal benchmark, Haiku 5.5 scored 1620 and Haiku 4.5 scored 735.
On the GDPVal benchmark, Opus 5.5 scored 1846 and Astra scored 1542.
Grok 4.7 achieved 71.0% on the Deep SWE software engineering evaluation, 38.0% on TerminalBench, and an Elo of 1,695 on the GDPVal evaluation, second only to Fable 5.1's 1,735.
All Facts (8)
GDPVal基准测试覆盖44个职业和9大行业，任务直接来源于真实职场场景
80%UnverifiedGDPval（General Domain Performance Validation）是一个综合性大语言模型评估基准，涵盖推理、知识问答、代码生成、多语言理解等维度
70%UnverifiedThe Tongyi Qianwen Flash edition starts at $0.03 and $0.13, and GLM scored 1,647 points on GDPVal.
50%UnverifiedThe new model achieved 1,620 points on GDPVal (compared to 735 for the previous generation), improved on OSWorld from 15.7% to 72.4%, on Terminal Bench from 0.0% to 39.2%, and on Humanity's Last Exam from 10.2% to 45.9% without tools and from 18.7% to 57.4% with tools.
50%UnverifiedAnthropic states that Haiku 5.5 outperforms GPT-6 Luna on certain metrics including GDPVal and Agentic Coding.
50%UnverifiedOn the GDPVal benchmark, GPT-6 Luna scored 1437, lower than Haiku 5.5's score of 1620.
50%UnverifiedOn the GDPVal benchmark, Haiku 5.5 scored 1620 and Haiku 4.5 scored 735.
50%UnverifiedOn the GDPVal benchmark, Opus 5.5 scored 1846 and Astra scored 1542.
50%