Anthropic Fable 5.2 首批实测:100+ 生成对比 Opus 表现
First impressions of @AnthropicAI’s Fable 5.2 are in. @petergostev puts it through 100+ generations,...
有人拿 Fable 5.2 跑了 100 多次生成,跟 Opus 在 3D 场景、小游戏、网站 UI 这些任务上正面对比。选模型纠结成本和质量的可以看看,比官方跑分实在多了。
Anthropic 的 Fable 5.2 迎来首批公开实测,petergostev 用 100+ 次生成任务对其进行了系统测试,并与 Opus 做直接对比。测试覆盖五类场景:3D 世界构建、可玩小游戏、网站 UI、数据可视化和动画。整套测试通过 LMArena 直播的形式完整呈现。
从测试设计看,这次对比聚焦在生成式前端和创意 coding 任务上,每个场景跑了大量样本,以观察模型输出的稳定性和质量分布,而不是单次最佳表现。结论涉及三方面:Fable 5.2 在哪些场景有提升、Opus 在哪些方面仍占优、以及两者的成本差异。
对开发者来说,这类大样本实测比官方跑分更接近真实使用体验,尤其在模型选择直接影响 token 成本的场景下,性能与价格的权衡数据有实际参考价值。
First impressions of @AnthropicAI’s Fable 5.2 are in. @petergostev puts it through 100+ generations,...
First impressions of @AnthropicAI ’s Fable 5.2 are in. @petergostev puts it through 100+ generations, comparing it with Opus across 3D worlds, playable games, website UI, data visualization, and animation. Where does Fable 5.2 improve, where does Opus still stand out, and at what cost? Watch the full breakdown youtube.com/live/QwaU4yU0P…📊 15 ⚡
This report is categorized under AITOP's "Models" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include Anthropic, Fable 5.2, Opus, 编程助手.
Source attribution: curated from lmarena.ai (@lmarena_ai).