Choose an AI model by testing your task and reading its official catalog. A newer name alone does not establish better results for your use case.
Read the catalog before the ranking
As of October 1, 2026, OpenAI lists GPT-6 Astra, GPT-6.1 Sol and GPT-6 Luna. Anthropic lists Claude Fable 5.1, Opus 5.5, Sonnet 5.5 and Haiku 4.5. Google lists stable Gemini 3.8 Flash and 3.5 Flash-Lite alongside Gemini 3.1 Pro preview. This is a dated selection, not an exhaustive ranking.
Check the exact model ID, release status, supported inputs and account availability. A chat application's features can differ from the underlying API. An alias can change; record the version used when reproducing an experiment.
Compare fairly
Our suggested exercise is to use five self-written questions with answers you can check. Keep the prompt and source material the same. Record correct answers, unsupported statements, response time and whether you had to repair the output. Include one question the supplied material cannot answer.
Choose the simplest model that meets your requirements. A benchmark score cannot replace your own task results, and our catalog summary is not a performance test.
Official sources
Recheck these pages before purchasing access or starting a project. See the dated models table for the same snapshot.