Published • loading... • Updated
New Anthropic, OpenAI Models Make Same Promise: A Little More for a Lot Less Money
Anthropic says the model cuts running costs by 40% on typical workloads and adds safeguards for cybersecurity and biology.
- Anthropic unveiled Claude Opus 5.5 on Tuesday, generating output over 30% faster than Opus 5 while cutting typical workload costs by around 40%. The model matches Claude Fable 5.1 performance on most tasks.
- The release follows CEO Dario Amodei's September 12 call to "pace the frontier" of AI development, urging the industry to slow capability rollouts to address safety concerns. Opus 5.5 is Anthropic's first major model since that announcement.
- Independent safety researchers Frontier Design and METR tested Opus 5.5 before release, with the model priced at $4 per million input tokens and $20 per million output tokens. Internal benchmarks show 66.4% on Terminal-Bench 4.0, outperforming competitors.
- OpenAI simultaneously launched lower-cost GPT-6 Sol and Luna models, intensifying pricing competition for enterprise developers. Fortune reports this commoditization of flagship capabilities could compress margins amid competition from low-cost open-weight alternatives.
- Anthropic plans to release Claude Sonnet 5.5 and Claude Haiku 5.5 in the coming weeks, expanding the new Claude 5.5 family. Reports suggest the company could seek a valuation approaching $2 trillion ahead of a potential initial public offering.
Insights by Ground AI
Podcasts & Opinions
130 Articles
130 Articles
Anthropic launches the offensive against GPT-6 Astra with Claude Opus 5.5, this is what it is capable of doing
OpenAI presents two new GPT-6 models that improve programming, reduce errors and significantly lower the cost of using artificial intelligence.
·Los Angeles, United States
Read Full Article
The fears that the AI escapes the understanding of its designers are growing.
·Montreal, Canada
Read Full ArticleCoverage Details
Total News Sources130
Leaning Left12Leaning Right12Center24Last UpdatedBias Distribution50% Center
Bias Distribution
- 50% of the sources are Center
50% Center
L 25%
C 50%
R 25%
Factuality
To view factuality data please Upgrade to Premium
