Grok 4.7 scores 35, near-bottom on Artificial Analysis as critic slams Grok Bot pricing
Hesamation · x · 2026-11-08
Hesamation argues Grok Bot's business side is ruining a well-engineered product: an Artificial Analysis chart shows Grok 4.7 (xhigh) scoring just 35, tied near the bottom with Tencent Hy4, far behind leaders like Sonnet 5.5 (max) at 55 and GPT-6.1/Terra (max) at 51.
He cites three issues: Grok 4.7 lagging in capability, tighter usage limits, and unattractive pricing. An app that locks users into one model with no cheaper/smaller option should be exceptionally smart or exceptionally cheap — Grok is currently neither.
He still uses Grok Bot daily for its reliability and simplicity, but says the pricing needs to catch up before anyone can justify it for coding/engineering work.
Related event: Grok 4.7 bottoms out in benchmarks as pricing draws criticism(2 posts)→
More from Models
- Ethan Mollick: AI now beats PhD experts at fact-finding and claim verification — emollick · 2026-11-08
- yacineMTB: Terra beats Opus 5.5 at getting real work done — yacineMTB · 2026-11-08
- Alibaba's Qoder launches stealth coding model Qoduck, free daily for individual users — alifcoder · 2026-11-08
- DeepSeek-V4.1-Flash-1017 praised for inference-efficiency-first design — Yuchenj_UW · 2026-11-08
- Labs are pacing themselves: the joke that Opus 5.5 may be the peak — adonis_singh · 2026-11-08
- Why LLMs abuse the em dash: corpus bias, autoregression and RLHF reward hacking — SparkyAI0815 · 2026-11-08