Home · Technology · Sep 28 archive
Anthropic Launches Cheaper, Faster Sonnet 5.5 AI Model
Confirmed
In Short: Anthropic has unveiled Sonnet 5.5, a new AI model designed to be a faster and cheaper alternative to its previous offerings. The company claims Sonnet 5.5 is 30% faster and up to 30% less costly than its predecessor, Sonnet 5.
Anthropic has unveiled Sonnet 5.5, a new AI model designed to be a faster and cheaper alternative to its previous offerings. The company claims Sonnet 5.5 is 30% faster and up to 30% less costly than its predecessor, Sonnet 5.
According to Anthropic, Sonnet 5.5 is optimized for well-scoped everyday tasks such as fixing bugs, creating polished documents, slides, and spreadsheets. The model is positioned as a complement to the more powerful Opus 5.5, which was released just last week.
In a statement, Anthropic noted that Sonnet 5.5 scored higher on agentic coding capabilities compared to Sonnet 5, though it did not surpass Opus 5.5 in all benchmarks. The model's performance on GDPval-AA v2.1, which measures real-world work across 44 occupations, was just two points behind Opus 5.5 and about 400 points ahead of Sonnet 5.
Despite the CEO's call for a slowdown in the industry to allow safety frameworks to catch up, Anthropic continues to release new models at a rapid pace. The company stated that Sonnet 5.5 does not advance the frontier of its model's capabilities and does not introduce new guardrails.
Sonnet 5.5 is also the first Sonnet model to beat Pokémon Red, working only from screenshots, according to Anthropic. The model's cybersecurity capabilities are comparable to those of Opus 5, and it will be subject to the same cyber safeguards.
Slack principal engineer Curtis Allen said Sonnet 5.5 outperformed Sonnet 5 on nearly all of Slack's offline Slackbot evaluations, finishing in fewer steps and with about 14% fewer output tokens. The input and output rates are half of Opus 5.5's.
As the AI model wars continue, Anthropic is pushing its AI business deeper into the cost-sensitive enterprise market with Sonnet 5.5, designed for routine coding and business tasks.
The rapid release of new models by Anthropic and other companies like OpenAI highlights the ongoing competition in the AI industry, with each company striving to offer more efficient and cost-effective solutions.
Anthropic's CEO, Dario Amodei, had previously called for the industry to 'pace the frontier,' but so far, this call has not produced any meaningful changes in the industry's pace.
While benchmarks capture only part of a model's abilities, Anthropic's latest release underscores the company's commitment to balancing rapid commercial expansion with growing scrutiny over AI risks.
What this adds
The rapid release of new models by Anthropic and other companies highlights the ongoing competition in the AI industry, with each company striving to offer more efficient and cost-effective solutions.
What's confirmed
- Anthropic said Sonnet 5.5's cybersecurity capabilities are comparable to Opus 5's.
What's still developing
- In this case, at least, we’re talking about the release of a new entry in the company’s mid-level Sonnet tier.
- The announcement did note, however, that the model was the least likely of all the models Anthropic had tested internally to actively search for vulnerabilities in its testing sandbox.
- Anthropic said Claude Haiku 5.5 will arrive in the coming weeks.
- Sonnet 5, 5.5’s predecessor, was announced about three months ago.
- Meta also announced a new model, which it said would power an upcoming feature associated with its smart glasses.
- In particular, Anthropic’s benchmarks show Sonnet 5.5 performing better than Opus 5.5 on agentic coding, likely because of its ability to spawn multiple agents without exceeding cost limits.