Artificial Analysis adds provider-level cache hit rates and Cost per Task

What happened

Artificial Analysis now publishes cache hit rates for tracked model endpoints and provider-level Cost per Task for open weights models. Cache hit rates are measured using multi-turn agentic coding tasks; provider cost comparisons use the same token counts from each model's Intelligence Index run. GLM-5.3 (max) costs vary ~9x, from $0.92 on Makora to $8.68 on Nebius, largely because of input pricing despite both having high cache hit rates. Across 21 providers (25 endpoints), cache hit rates range from 41% to 100%.

The story so far

Artificial Analysis now publishes cache hit rates for tracked first-party and third-party model endpoints, along with provider-level Cost per Task comparisons for open weights models. It measures cache hit rates by running multi-turn agentic coding tasks and reading cached-token counts from each provider’s usage response. The cost comparisons apply each provider’s input, output, cache hit and cache write prices at its measured cache hit rate, using the same token counts from each model’s Intelligence Index run. Artificial Analysis reports that GLM-5.3 (max) costs about nine times as much per task on Nebius ($8.68) as on Makora ($0.92), largely because of input pricing despite both providers having high cache hit rates. Across 21 providers covering 25 endpoints, measured cache hit rates range from 41% to 100%.

Keep reading

Trending now

1. 1Meta Releases Muse Spark Open With Mac and Cloud Support, Topping US Open-Weight Index
2. 2Google Launches Gemini 4 Argon Globally, Expands Paid Access as Benchmark Results Vary
3. 3OpenAI Updates GPT-6 Astra Pro in ChatGPT With Parallel Agents and Faster Responses
4. 4Meta’s New Business Terms Raise Dispute Over Website Crawling and AI Training Rights