xAI Launches Grok 4.5: 1.5T Parameters, Focus on Coding & Agents, Low Price
Decision Brief
What changedxAI releases Grok 4.5, a frontier model specialized in coding and agentic tasks, emphasizing cost-effectiveness over benchmark performance.
Why it mattersFor API developers, 1.5T parameters at $2/$6 per 1M tokens offers exceptional value.
Who should careTeams building on model APIs, AI coding tool users
Affected stackCursorxAI
Builder actionUpdate Cursor to the latest version
Source confidenceMedium · Reliable media or first-hand reporting
xAI officially launched Grok 4.5, a frontier model with 1.5 trillion parameters (1.5T)—three times that of Grok 4.3. Elon Musk calls it 'Opus-level' but faster, more token-efficient, and cheaper. Pricing: $2 per 1M input tokens, $6 per 1M output tokens, with cache hit discounts. Context window is expected to return to 1M tokens soon. Cursor co-trained Grok 4.5 and calls it the most powerful model yet, extending beyond pure software engineering. Early ecosystem support includes Grok Build/API, Hermes Agent, Portal, and OpenRouter. Developers on these platforms can immediately test Grok 4.5's agentic and coding capabilities.
Summary basis: official / RSS sourceCompiled from the source scope noted above; the original remains authoritative.
Sources
- AINews（smol.ai）
Weekday recap of top news for AI engineers — dense, builder-oriented signal.
- AINews（smol.ai）
留言
登入后即可留言，和其他 builder 交换实测心得。
还没有留言，抢头香。