21 stories tagged by Digg
Technology
The New Stack says Anthropic claims Opus 5.5 is 40% cheaper and 30% faster than Opus 5. On hard reasoning tests, the savings held up, but the speed claim fell short.
AI
A user says Anthropic attributes potential savings of up to 30% on most jobs to the model needing fewer tokens. They also cite a 70.6% Terminal-Bench 4.0 score, up from 10.3%.
AI
Theo ranks 15 AI models with Fable 5 in S+ and Opus 5 in D.
AI
Anthropic shared results from testing Claude on de novo protein binder design.
AI
Suggestion follows complaints that high effort causes overcorrection on minor comments.
AI
Tech creators share disappointment with Opus 5's overzealous fixes and basic errors.
AI
Social media users and researchers question Opus 5 performance and Fable model choices.
AI
Fireworks AI benchmark compares Kimi K3 and Opus 5 on SWE and coding tasks using task-based pricing.
AI
AI practitioners report Opus 5 underperforms in practice despite benchmark wins over Fable.
