New frontier chart puts Grok 4.5, SWE-1.7 and Opus 5 on one cost curve
GavinSBaker · x · 2026-07-25
The new cost–performance frontier now includes Grok 4.5, SWE-1.7 and Opus 5
A shared benchmark chart shows the latest frontier models clustering on a new Pareto curve: Grok 4.5, SWE-1.7, and Opus 5. The post argues that “intelligence per dollar” will become the key metric over time.
The attached chart compares score vs. rollout cost and suggests that models like Opus 5 are pushing the frontier upward while staying in roughly the same cost band as prior top-tier systems. It also notes that K3 could join the frontier once it becomes available on inference clouds.
More from Models
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11
- Opus Refuses Protein Research Codebase Over 'Safety' Concerns, Dev Considers Rolling His Own — josephdviviano · 2026-09-11
- User Hails Unconfirmed 'DeepSeek 4.1 Flash' as an Inflection Point in LLMs — himanshustwts · 2026-09-11
- Terminal Bench v4: GLM-5.3 Leads at 41.9%, Kimi-K3 Underwhelms at 12.6% — Ok_Warning2146 · 2026-09-11
- GPT-6 Astra beats Factorio with enemies in 44 in-game hours at ~$4,500 API cost — liminal_bardo · 2026-09-11