OpenAI announced GPT-Red, an internal model for finding prompt-injection vulnerabilities at scale.
> GPTβRed is a strong red-teamer, and our previous models are highly vulnerable to its prompt injection attacks.
> We use GPTβRed to adversarially train GPTβ5.6, making it much more robust to prompt injections.
> GPTβRed is a strong red-teamer, and our previous models are highly vulnerable to its prompt injection attacks.
> We use GPTβRed to adversarially train GPTβ5.6, making it much more robust to prompt injections.
β€5π2 1
Thinking Machines debuts open-weight Inkling AI model
Thinking Machines Lab released Inkling, an open-weights multimodal MoE model with 975B parameters, 1M-token context, and tunable reasoning. It supports customization, broad deployment options, and strong coding, reasoning, vision, and audio results.
π #ai @testingcatalog
Thinking Machines Lab released Inkling, an open-weights multimodal MoE model with 975B parameters, 1M-token context, and tunable reasoning. It supports customization, broad deployment options, and strong coding, reasoning, vision, and audio results.
π #ai @testingcatalog
TestingCatalog AI News
Thinking Machines debuts open-weight Inkling AI model
Thinking Machines Lab launches Inkling, its open-weight Mixture-of-Experts AI model, offering broad customization and multi-modal support.
π₯4β€2π2
This media is not supported in your browser
VIEW IN TELEGRAM
Kimi K3 has been teased officially π
β€8π5π₯2
Anthropic reset 5h and weekly rate limits on Claude for all users.
3 more Fable 5 days left (until July 19) and it likely wonβt come back.
Last testing chance π
3 more Fable 5 days left (until July 19) and it likely wonβt come back.
Last testing chance π
β€10π8π3 2
Early look at Kimi K3 generations from Moonshot AI on Arena
Moonshot AI appears close to launching Kimi K3, with leaks, teasers, and test sightings pointing to an imminent debut. Early reports suggest frontier-level performance, strong 3D and coding output, but slower runtimes on complex tasks.
π #kimi @testingcatalog
Moonshot AI appears close to launching Kimi K3, with leaks, teasers, and test sightings pointing to an imminent debut. Early reports suggest frontier-level performance, strong 3D and coding output, but slower runtimes on complex tasks.
π #kimi @testingcatalog
TestingCatalog AI News
Early look at Kimi K3 generations from Moonshot AI on Arena
Moonshot AIβs Kimi K3 launch seems imminent, with leaks and teasers hinting at a groundbreaking multi-trillion-parameter MoE model.
β€6π3
MOONSHOT π₯: Kimi K3 from Moonshot AI landed on the first place on Frontend Code Arena with 1679 points, outperforming Claude Fable 5 and GPT-5.6 Sol by far.
β€12π₯3π2
Meta β€οΈ OpenRouter
Muse Spark 1.1 is now available on OpenRouter for US users.
Muse Spark 1.1 is now available on OpenRouter for US users.
β€5π3
This media is not supported in your browser
VIEW IN TELEGRAM
Google is working on a native menu for Skills on Gemini desktop. These Skills will potentially become available in all chats (and not only in Spark as it is now).
Users will be able to upload, create and edit their Gemini Skills, as well as use Gemini to create them.
Skill folders will be available there too, allowing users to select local folders containing necessary Skills.
Users will be able to upload, create and edit their Gemini Skills, as well as use Gemini to create them.
Skill folders will be available there too, allowing users to select local folders containing necessary Skills.
β€8π4
ICYMI π: Grok Heavy subscribers now get X Premium+ for free!
Would be cool for X Premium Business subscription to include Grok Heavy one day.
Would be cool for X Premium Business subscription to include Grok Heavy one day.
β€5π4
OPENAI π₯: ChatGPT desktop app now includes ChatGPT chats and projects too! The history is being synced across desktop, web, and mobile.
The Chat and Work mode switcher now works consistently with the web version, and Codex mode continues to function as before.
A very nice change! π
The Chat and Work mode switcher now works consistently with the web version, and Codex mode continues to function as before.
A very nice change! π
π₯10π4β€3
X now has a dedicated AI timeline to follow latest AI model releases and more!
To enable it, go to Add+ > Add Timeline > Select βArtificial Intelligenceβ > Profit!
Even though my βFor Youβ timeline is exactly the same, βAIβ is now my default.
Cya there π
To enable it, go to Add+ > Add Timeline > Select βArtificial Intelligenceβ > Profit!
Even though my βFor Youβ timeline is exactly the same, βAIβ is now my default.
Cya there π
1β€11 4π2