Grok 4.5 Leads Frontier Models in Speed
Daniel_Farinax · x · 2026-07-15
Grok 4.5 is described as significantly outpacing GPT-5.5 Sol and Claude Fable 5 in speed:
- Throughput is around 80–112 TPS, compared to roughly 55–70 TPS for the Claude series.
- First token latency is under 1 second.
- For coding and agent tasks, overall end-to-end time is typically 2x faster.
- Processing time is halved, and it consumes fewer tokens.
The original post concludes that it is the reigning "speed king" among frontier models.
Related event: Grok 4.5 Draws Praise for Speed Lead(2 posts)→
More from Models
- A 2-minute Astra audit at low setting wiped a Plus user's full 5-hour limit — Existing-Slide7395 · 2026-09-07
- DeepMind-Princeton paper shows LLMs causally use confidence to decide whether to answer — GoogleDeepMind · 2026-09-07
- Qwen 3.8 Next Flash is painfully verbose: 13-minute thinking on single coding prompts — Infinite-Local5435 · 2026-09-07
- Philosopher asks GPT-6 to review his Oxford book: result rivals top-journal reviews — anselm · 2026-09-07
- Leaker claims xAI is preparing Grok 4.7, hints at another surprise — mark_k · 2026-09-07
- Local LLMs now near Opus-level — what's still keeping them behind closed models? — mrsalvadordali · 2026-09-07