AI News
AI community posts, news and release updates
Latest updates
A new tool that automatically suspends LLM agents before hitting rate limits and resumes them cleanly afterward.
- LLM agents
- tool
- rate limits
Google releases three new Gemini models: 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, targeting various use cases.
- Gemini
- Flash-Lite
- Flash Cyber
- News
Grok gets major upgrade in Tesla 2026 Summer Release: phone calls, music, climate control
Published Jul 21, 2026Tesla's 2026 Summer Update enables Grok to make phone calls, search and play music, adjust climate, open glovebox, and answer Tesla-related questions via natural voice.
- Grok
- Tesla
- Update
- voice commands
Google launched Gemini 3.6 Flash, with output price dropping from $9.00 to $7.50 per million tokens, outperforming 3.1 Pro in nearly all benchmarks, competing with GPT-5.6 Luna and Sonnet 5 as a solid mid-tier model.
- 基准测试
- 性价比
- Gemini 3.6 Flash
- News
Tesla 2026 Summer Update unveiled: Grok calls, personalized navigation, self-driving stats
Published Jul 21, 2026Tesla's Summer 2026 update includes Grok phone calls, route prioritization by history, self-driving stats from app, automatic navigation to frequent places, and more.
- Tesla
- 2026
- features
- software update
Google launched three new AI models: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber, expanding its lineup.
- Gemini 3.6 Flash
- Gemini 3.5 Flash-Lite
- Gemini 3.5 Flash Cyber
Google unveiled new Gemini models and announced it has begun pre-training Gemini 4, the next generation.
- AI模型
- Gemini 4
- News
Pricing Announced for Three New Gemini Models, Gemini 4 Pre-training Underway
Published Jul 21, 2026Google announced pricing: Gemini 3.6 Flash ($1.50 in, $7.50 out), 3.5 Flash-Lite ($0.30/$2.50), and 3.5 Flash Cyber. Gemini 3.5 Pro still in testing, Gemini 4 pre-training has begun.
- 定价
- Gemini 4
- Gemini 3.6 Flash
A new open-source diffusion-based ASR model claims 15x speedup over Whisper, offering a fresh alternative for speech recognition.
- 开源
- 扩散模型
- Whisper
- ASR
- News
NVIDIA Blackwell Ultra Sets World Record for DeepSeek-V3 Pre-Training: 1,648 TFLOPs per GPU
Published Jul 21, 2026NVIDIA's Blackwell Ultra achieves 1,648 TFLOPs per GPU on DeepSeek-V3 671B pre-training, 3x the previous generation, thanks to co-design and continuous software optimization across Megatron-Core, TorchTitan, and JAX.
- Nvidia
- DeepSeek-V3
- 训练性能
- Blackwell Ultra