Frontier Daily (Aug 14): 11 items from DeepMind, OpenAI, Latent Space and more + X supplements

Every day Laojin pulls the official channels and compiles the past 48 hours of AI news into a short brief: curation only, no commentary, with the original link and publish time attached to every item.

1. Introducing Gemini 3.7 Flash

Google DeepMind introduces Gemini 3.7 Flash on its blog.

Source: Google DeepMind Blog · 2026-08-13 · Original

2. Previewing Ultrafast mode: GPT-5.6 Sol at up to 14X the speed

OpenAI previews Ultrafast, a new API tier powered by Cerebras that runs GPT-5.6 Sol up to 14x faster with up to 750 output tokens per second.

Source: OpenAI Blog · 2026-08-13 · Original

3. OpenAI appoints Dali Rajic as Chief Revenue Officer

OpenAI appoints Dali Rajic as Chief Revenue Officer to lead its global revenue organization.

Source: OpenAI Blog · 2026-08-13 · Original

4. [AINews] SpaceXAI Grok 4.6 and Grok @Bot

SpaceXAI (the former Cursor team) ships Grok 4.6 and Grok @Bot, a knowledge-work model praised as arguably the second best globally and top by efficiency.

Source: Latent Space · 2026-08-13 · Original

5. DeepSeek V4 Pro 0813 (on OpenRouter)

DeepSeek V4 Pro 0813 is now available via API, with notable output differences across reasoning levels and possible open-weight release hinted by prior releases.

Source: Simon Willison's Weblog · 2026-08-12 · Original

6. alchemy-utils 0.1a0

Simon Willison releases alchemy-utils 0.1a0, a database-agnostic SQLAlchemy-based library built with Codex and GPT-5.6 Sol Ultra as an alpha.

Source: Simon Willison's Weblog · 2026-08-12 · Original

7. Google DeepMind Releases Gemini 3.7 Flash

Google DeepMind launches Gemini 3.7 Flash, significantly stronger in coding, knowledge work, and web development; it jumps to #8 on Code Arena WebDev (from #19) and improves rankings across Text Arena categories, now available in Gemini app and APIs.

Source: X (Grok x_search, @GoogleDeepMind) · within 48h · Original

8. Elon Musk Announces Grok 4.7 Coming in 3-4 Weeks

Elon Musk reveals Grok 4.7 has completed initial training and is incorporating massive SpaceX data, expected in 3-4 weeks; it will be significantly better than 4.6 and excel in real-world engineering over all current models.

Source: X (Grok x_search, @elonmusk) · within 48h · Original

9. Grok 4.6 Achieves SOTA on Databricks OfficeQA Pro V2

xAI's Grok 4.6 reaches state-of-the-art on Databricks' latest OfficeQA Pro V2 benchmark, excelling in document understanding and data reasoning, hailed as the highest intelligence-per-dollar AI model.

Source: X (Grok x_search, @elonmusk) · within 48h · Original

10. Arena.ai Leaderboards Updated: Gemini 3.7 Flash Dominates

Arena.ai community evaluations show Gemini 3.7 Flash (High) scoring 1588 in Code Arena WebDev and 1490 in Text Arena, topping multiple categories and redefining the Pareto frontier with big gains over its predecessor.

Source: X (Grok x_search, @arena) · within 48h · Original

11. Elon Musk: Grok 4.6 is the Most Efficient High-Intelligence AI

Elon Musk affirms that Grok 4.6 offers the highest intelligence per dollar and is the most efficient high-intelligence AI, while highlighting its humor and real-world task capabilities.

Source: X (Grok x_search, @elonmusk) · within 48h · Original

---

_Sources: official blogs of OpenAI, Google DeepMind and Qwen; official RSS of Simon Willison, Latent Space, Import AI and One Useful Thing; plus change monitoring of anthropic.com. X sources are retrieved live via Grok x_search with per-link verification. This is a source digest — no commentary from Laojin._