Expired50% confidenceFactExact time
Musk claims that Grok 4.5 performs comparably to competing products but is faster, is trained on real interaction data from Cursor engineers, and uses half the token volume.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/12/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/10/2026(expired)
Sources
Grok 4.5 Launch: Coding AI Market Reshuffle, Anthropic Claims Half the Territory
bilibili竹言见智7/9/2026
Related Claims
UnverifiedGrok 4.5 is trained on real interaction data from Cursor engineers, uses half the tokens, and has an overall cost of less than half that of competing products.81% similarUnverifiedGrok 4.5 reportedly introduces more Chain-of-Thought reinforcement training.69% similarUnverifiedMixed-precision training can typically increase training speed by 1.5–2 times and reduce VRAM usage by nearly half.69% similarUnverifiedTechniques such as mixed-precision training and gradient accumulation enable training of larger models under limited GPU memory.68% similarUnverified混合精度训练在不显著损失模型精度的前提下，可将训练速度提升约2倍，显存占用降低近一半67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/490338API
curl https://kongchang.com/api/v1/knowledge/claims/490338MCP
get_claim(id=490338)