Product Launchbenchmarksmlopsllmopen source
AI Energy Score Updates Leaderboard, Benchmarks Reasoning Costs
9.0
The AI Energy Score today launched a refreshed leaderboard and benchmarking package, adding 39 models and a dedicated reasoning task. The update reports reasoning modes use about 30× more energy on average, with some models showing 150–700× increases, and finds mixed efficiency changes across 15 comparable models. The project released an open-source AI Energy Benchmarks package and Salesforce has adopted energy-transparent model cards.
Key Points
- 1Quantifies that reasoning modes increase energy use about 30× on average, up to 700× in examples
- 2Shows mixed efficiency trends: 9 of 15 comparable models consumed equal or more energy than predecessors
- 3Encourages standardized benchmarking and routing strategies; Salesforce adopted energy transparency in model cards
Scoring Rationale
Official benchmarking reveals large energy costs from reasoning modes, but update is incremental rather than paradigm-shifting.
Sources
Public references used for this report.
Practice with real Logistics & Shipping data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
High-Value Overnight OrdersEasyDelivered International ShipmentsMediumOn-Time Delivery Rate by CarrierHard
250 free problems · No credit card
See all Logistics & Shipping problems
