August belonged mainly to OpenAI, but much more happened. We saw GPT-5, open-source models, and the new Codex IDE. Anthropic added a million-token context and Google showed two new models for working with images.
GPT-5
OpenAI launched a new generation of its GPT-5 model. What's new?
- Multiple versions for different needs: standard, mini, and nano. Mini and nano are tuned for speed and price; standard offers balanced performance.
- Cheaper to run: a router splits queries between models and uses resources more efficiently.
- Less people-pleasing: GPT-5 cuts down the over-accommodating "sycophancy", reads advanced questions better, and adapts its communication style.
- Focus on technical roles: better results on expert and programming tasks, even if a solution can take longer.
- Integration: access to external services (GitHub, Google Drive, Gmail, Calendar) for Plus users.
- Reasoning effort: you can set how "deeply" the model reasons.
- Verbosity: you control the length and detail of the answer, from concise to explanatory.
- Custom tools: tools defined in plain text or with a context-free grammar, for more flexible tool calling.
GPT OSS
OpenAI also introduced open-source models.
- Two models with 20M and 120M parameters, with fairly different architectures.
- For their size, both deliver very good results.
- They use Mixture of Experts, so only part of the model activates per query.
- They emphasize sustainability and saving resources.
GPT Real time
The next release from OpenAI is a model tuned for live scenarios, especially call centers.
- Real-time speech-to-speech.
- Handles image inputs too.
- Can call tools through MCP.
OpenAI Codex IDE
OpenAI is expanding its portfolio for developers as well.
- A new coding assistant right in the IDE.
- Available under the single OpenAI license.
- Handles full agent mode for vibe coding, like Claude Code.
Claude 4 Sonnet
Anthropic made a major move: Claude 4 Sonnet now handles 1 million tokens in context. That means you can work with a huge amount of text at once. It shows up in the price too.
Gemini 2.5 Flash Image
Google is out with a new image-editing model. A few days before its official release, it showed up on one of the community platforms under the name Nano Banana.
- It focuses mainly on photo editing. It can color a black-and-white shot or work out how an object would look from behind.
- It converts 2D to 3D, merges several images, or adds realistic shadows.
- It still struggles with text inside images, but otherwise it's among the best we've seen so far.
Here's an example of its work, with the prompt: "This guy sitting with that dog on the sofa in the same room."
Claude for Chrome
Anthropic is testing a Chrome extension.
- An agent assistant right in the browser.
- It can order goods, manage emails, and work with the web.
August brought major news mainly from OpenAI: GPT-5, open-source models, and a new IDE. Anthropic moved to a million-token context, Google to more realistic image work. The common thread is more emphasis on practical use and more control for developers, who now get tools better suited to real-world scenarios.
Want to stay one step ahead?
Don't miss our best insights. No spam, just practical analyses, invitations to exclusive events, and podcast summaries delivered straight to your inbox.