Compare models
Same data as Global rankings; composite score uses the same weighting as the Multimodal model leaderboard.
Recommended score follows the same site weighting (context first, then blended price), with scenario-based presets.
Data sources & methodology
Source: OpenRouter
Metrics: context_length_desc + avg_price_prompt_completion
Update cadence: daily
Fetched at: 2026-06-24T04:09:26.639Z
Method: webx-ranking-v2
Pick 2-3 models quickly
Starter trio: OpenAI: GPT-4o
Pick 2–3 models on Global rankings first, or use the link below.
FAQ
How is the composite score calculated?
It uses the same on-site weighting as the Multimodal model leaderboard: context window first, then blended 1M-token price. It is not LMSYS Chatbot Arena ELO.
Can I compare more than three models?
The interactive table supports up to three models at once. Use query parameters or the model chips to adjust your selection.
How often is pricing refreshed?
Snapshots refresh on a scheduled pipeline cadence (typically every six hours). Check the disclosure block for the latest fetched timestamp.