Skip to content
Trending storyDeveloping

Artificial Analysis publishes cost per task for GLM-5.3 (max) across providers

2 articles1 sourcesince Oct 13Last article 1h ago ·

Overview

AISummary of 2 articles

Artificial Analysis publishes cost per task by provider for GLM-5.3 (max), with a comparison hosted on its site.

The site reports costs from $0.92 on Makora to $8.68 on Nebius, a roughly 9x spread that the firm attributes mainly to differences in cached input pricing. It also reports cache hit rates ranging from 41% to 100% across 21 providers and 25 endpoints.

Cost per task is calculated from each provider's input, output, cache hit and cache write prices, applied at its measured cache hit rate. Cache hit rate is the share of cacheable input tokens served from cache.

Written by AI from the articles below · updated Oct 13, 3:30 PM ET

Check the sources:

Latest developmentOct 13, 3:17 PM ET
Artificial Analysis publishes cost per task for GLM-5.3 (max) providers

Developments

2 developments

  1. Oct 13, 3:17 PM ET · 1 article
    Artificial Analysis links cost-per-task comparison for GLM-5.3 (max) providers
    Artificial Analysis: Artificial Analysis publishes cost per task for GLM-5.3 (max) providers
  2. Oct 13, 3:17 PM ET · 1 article
    Artificial Analysis publishes cache hit rates and Cost per Task by provider for open weights models
    Artificial Analysis: Artificial Analysis shows GLM-5.3 cost per task varies 9x across providers

Article timeline

The articles in this story. Times are ET.

Oct 13
  1. Artificial AnalysisOfficial
    Artificial Analysis publishes cost per task for GLM-5.3 (max) providers

    AIArtificial Analysis links to a comparison of cost per task across all GLM-5.3 (max) providers. The post contains no figures, and the comparison is hosted on the Artificial Analysis site.

  2. Artificial AnalysisOfficial
    Artificial Analysis shows GLM-5.3 cost per task varies 9x across providers

    AIArtificial Analysis now publishes cache hit rates and cost per task by provider for open-weights models. It reports GLM-5.3 (max) costs from $0.92 on Makora to $8.68 on Nebius, mainly because of differences in cached input pricing. Cache hit rates range from 41% to 100% across 21 providers and 25 endpoints.

    Image from @ArtificialAnlys's post

Heat trend

Not enough continuous observations to show a trend yet.