Musk caps Tesla AI rivals, directs staff to his own Grok
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Elon Musk told Tesla staff to switch to Grok "when possible" via a Friday memo, citing lower token costs and asking engineers to email him feedback directly
- Tesla capped employee AI spending at $200 per week earlier in the week, applying the limit to Anthropic, OpenAI, and Google models but exempting xAI's Grok
- Grok 4.5 ranks 9th overall (76.3) on an aggregate multi-domain leaderboard, behind multiple OpenAI, Anthropic, and Google models, with the lowest coding score (68.6) on the board
- Musk conceded on X that "Fable is definitely better than Grok 4.5," yet pitched the model as "Opus-class" at roughly $0.13 per task versus $1.57 for Claude Fable 5
- xAI released Grok 4.5 jointly with Cursor, the coding startup SpaceX is acquiring at a $60 billion valuation; Cursor disclosed an earlier snapshot of its codebase was accidentally included in Grok 4.5's training data, inflating one headline coding score
- Tesla engineers broadly prefer Anthropic's Claude for day-to-day development work, according to four people familiar with internal usage, despite months of internal Grok testing
Why it matters: Tesla engineers were already choosing Claude on the merits, yet Musk is using a spending cap that conveniently exempts his own product to redirect them toward a model he admits is worse — forcing a self-dealing trade-off that overrides engineering judgment with a price advantage that exists because Grok underperforms.
