Quadrillion says Anthropic’s Opus 5 is faster than Opus 4.8 on hard ML workloads
igarciacamargo · x · 2026-07-25
Quadrillion AI says it tested Anthropic's Opus 5 on hard research-science and ML workloads and found it meaningfully faster than Opus 4.8 while matching or beating it on quality.
- The attached QBench chart shows Opus 5 with a median wall-clock time of 59m50s versus 1h52m for Opus 4.8.
- The post frames the result as a real workload win rather than just a benchmark bump.
More from Models
- Descript ships Claude Opus 5 on day one inside its video editor — descript · 2026-07-25
- Opus 5 gets an unusually strong thumbs-up for deep-research workflows — madhavsinghal_ · 2026-07-25
- CogSci talk links linguistic efficiency to how LLMs learn meaning — ChrisGPotts · 2026-07-25
- A meme turns the Opus 5 benchmark into an OpenAI battlefield stare-down — kimmonismus · 2026-07-25
- OpenAI says GPT-5.6 cut one production agent’s build time 2.2× and cost 27% — OpenAI · 2026-07-25
- Opus 5 reportedly scores 42/42 on IMO 2026 without tools — Afinetheorem · 2026-07-25