Anthropic launches Claude Opus 5.5, cheaper than Opus 5
Anthropic released Claude Opus 22 on 5.5 September, the first model in the Claude 5.5 family. According to Anthropic, it performs at the level of Claude Fable 5.1 on most work and costs 40% less than Opus 5 on typical workloads. The API price is $4 per million input tokens and $20 per million output tokens (Opus 5: $5 and $25); the model is available via claude.ai, the API, AWS, Google Cloud and Azure.
What does this mean? If you use Claude via the API or a cloud platform, you now get more quality for less money, and it's worth switching your settings to claude-opus-5-5.
OpenAI introduces GPT-6 Sol and Luna with lower API prices
OpenAI has launched two new GPT-6 variants: Sol ($2 per million input tokens, $10 per million output tokens) and the lighter Luna ($0.10 and $0.50). Both are available to Plus, Pro, Business, Enterprise and Edu users and via the API as gpt-6-sol and gpt-6-luna. GPT-6 Astra remains the most powerful model, at $10 and $50 per million tokens; in addition, improved prompt caching gives up to 90% off repeated input tokens.
What does this mean? For many everyday tasks you no longer need the most expensive GPT-6 model: Sol or Luna can noticeably lower your API costs.
AI cameras in shops: the EU AI Act requires no notification, but the GDPR does
Shops that use AI to analyse camera footage to flag possible shoplifters need not disclose this under Article 50 of the EU AI Act, according to ICT lawyer Arnoud Engelfriet: that transparency obligation applies only to AI that communicates directly with people, such as a chatbot. Under the GDPR, however, the processing must be clearly disclosed, at a minimum in the privacy statement and probably also with a sign in the shop. It remains open whether such flagging falls under the prohibition on automated decision-making (Article 22 GDPR).
What does this mean? If you deploy AI on camera footage or other personal data, look not only at the EU AI Act but also at your information obligations under the GDPR.
Google releases Gemini 3.8 speech models with watermarking
On 23 September, Google introduced Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS, two text-to-speech models being rolled out in the Gemini API and Google AI Studio; business access via Gemini Enterprise will follow. Google cites support for more than 100 languages and dialects. Every generated audio clip receives an invisible SynthID watermark.
What does this mean? If you build voice agents, audio versions or voice-overs, there is a new speech model to test; do check for yourself how natural the Dutch sounds.
