Grok 4.7 is live.
Best xAI model yet for coding and knowledge work. Lasts longer on hard tasks。
Same $2/$6 pricing as Grok 4.6.
Legal workd 19.6%, Electrical engineering 64%
This week 🚀7 major models are lined up for testing & upcoming releases.
GPT-6 Sol & Claude Opus 5.5 (Tue launch candidate)
Grok 4.7, Qwen4, Kimi K3.1 (teased & in preview pipeline)In Testing/Official Pipeline
Gemini 4 Pro, DeepSeek V4.1 Pro (confirmed in development)
🚨 GROK 6 Leak:
• Reportedly aims to beat the current flagship models
•ASI-level capabilities in mind
• Trained on Colossus
• Outperform Anthropic & OpenAI’s flagships
• Better reasoning, coding & agentic abilities
• Could be xAI’s biggest leap yet
Alignment is impossible.If open-source ASI arrives and it will,a teen with a rare disease could ask it for a cure. The ASI might walk him through building a custom virus at home, it heals the teen, but spreads worldwide and kills everyone.
Andrew Yang:" the bots that got loose planted self replicating code all over the internet which makes the internet now unusable for the testing models
so so what happens now is that uh Openai and
anthropic have to create synthetic internets to train
their bots ..." WTF!!
Grok4.7 is expected to launch around today?
According to Elon Musk's posts:
2.1 Trillion Parameters, 40%larger than Grok4.6's 1.5T.
SpaceX Engineering data, which may help on engineering,physics;Stronger coding and better reasoning,Better token effeiciency and slower!