Grok Voice Is Live on fal for Low-Latency Voice Agents
Grok Voice is now available on fal, with 0.70-second response time, tool calls that finish before the sentence ends, transcription, multilingual TTS, and voice…
Global AI Digest
-
-
Union Alpha Gets Speed Improvements: Try It Again
OpenCode says it has made improvements to Union Alpha after users reported slow performance, and encourages users to retry the model. More optimizations are sai…
-
GPT Images 2.5 Now Available on GMI Cloud
GPT Images 2.5, described as the #1 text-to-image and image editing model, is now available on GMI Cloud. Users can start creating immediately.
-
Midjourney Announces Weekly Office Hours for September 16
Midjourney's official account announced its Weekly Office Hours for September 16, sharing a link to the session. The announcement follows recent community polls…
-
OpenRouter Spotlights Union Alpha: Free 256K-Context Multimodal Model for Research, Coding, and Agentic Workflows
OpenRouter highlights Union Alpha, a free multimodal model with 256K context and tool calling, aimed at research, coding, and agentic workflows, with a demo sho…
-
Grok Build Now Remembers Your Project Conventions and Decisions Across Sessions
Grok Build introduces persistent memory that learns from your usage, plus /memory and /dream commands to browse and organize project knowledge.
-
OpenCode2 Lets You Continue Subagent Work and Pick Smarter Models for Reviews
OpenCode2 adds subagent continuation and per-subagent model selection, letting developers use smarter models for reviews and continue implementing in the same s…
-
GPT-6 Astra Tops Updated BrokenArXiv and ArXivMath Benchmarks
Updated BrokenArXiv and ArXivMath benchmarks focus on ArXiv conjectures refuted in the past month, with models evaluated inside a harness. GPT-6 Astra tops the…
-
Faxie Launches: AI Video Generation Inside Slack, Powered by MiniMax H3
Faxie, a new app by Tom Shapland, brings AI video generation directly into Slack, using MiniMax H3 for realistic, prompt-accurate videos to help hybrid teams co…
-
Preview: Faster App Testing with jev + opencode Browser-Use CLI
An early preview shows how fast browser automation can be for testing apps, combining Typesafe AI's jev with opencode's browser-use CLI.
