How much information the model can process in a single conversation (measured in tokens; ~750 words per 1K). More context = it can read longer documents without forgetting the beginning.
$/M input
What it costs to send one million tokens to the model (your questions, documents or instructions). Shorter bar = cheaper.
$/M output
What it costs for the model to generate one million tokens of response. It's usually pricier than input; it matters a lot if you generate long texts.
Modality
What kinds of data it accepts and produces. Example: text+image→text means it can receive text and images and responds with text.
// Source: OpenRouter public catalog. Prices in USD per million tokens.
// Reasoning, speed and cost benchmarks — not just a catalog · Last updated: 17 oct 2026, 02:58
Intelligence
Artificial Analysis Intelligence Index · higher is better
Speed
Output tokens per second · higher is better
Cost per Task
Average cost (USD) per index task · lower is better
Model
Provider
Intelligence
Code
Speed (tok/s)
$/M input
$/M output
Intelligence
Composite index from Artificial Analysis (reasoning, general knowledge, math). Higher = better answers on complex tasks.
Code
Performance specifically on programming tasks. Useful if you're looking for a model to write or debug code.
Speed
Response tokens generated per second (median). Higher = faster responses, important for real-time chat.
// Source: Artificial Analysis. Reference indexes; they don't guarantee results for your particular use case.
// AI applied to business
Want to apply AI to a real process in your company?
It's not about picking the trendiest model. Let's review the process, the available data and the result you're after before deciding.