Changelog
What changed, week by week.
Pricing updates, new calculators, methodology revisions. Updated every Monday. If you spot a stale rate, we usually fix it within 48 hours.
- Launch
SmartCloudSuites enters public beta
First public release. Six categories scoped, five calculators live, ten more in active build. The LLM API Cost Calculator is the flagship — token-level comparison across 14 frontier models with prompt-caching modeled by default.
- Launch
Brand identity finalized
Olive green + antique brass palette across the site. Premium operator aesthetic — designed to read as a serious research publication, not an SEO farm.
- Pricing
Anthropic Claude Sonnet 4.6 — pricing verified
Confirmed $3.00 / $15.00 per million input/output tokens. Cached input at $0.30/M (10% of base, 5-minute TTL). No changes from prior week.
- Pricing
OpenAI GPT-4o mini — pricing verified
Confirmed $0.15 / $0.60 per million input/output tokens. Cached input at $0.075/M. Still the cheapest first-party frontier-class option for high-volume English text workloads.
- Pricing
DeepSeek V3 — pricing verified
Confirmed $0.27 / $1.10 per million input/output tokens. Off-peak discount of 50% applies between 16:30–00:30 UTC. For batch-tolerant workloads, this is the cheapest frontier-class option indexed.
- Methodology
Prompt caching modeled by default
Added cached-input ratio slider to the LLM cost calculator. Reflects the reality that most production workloads have 30–60% repeated input (system prompts, knowledge bases). Defaults to 0.30 — adjust per your workload.
Want pricing updates pushed to your inbox? Subscribe via the newsletter form on the home page.