Kimi K3.1 scores 55 on Artificial Analysis, placed on the Pareto line

What happened

Artificial Analysis, a company that tests AI models, posted results for Kimi K3.1 Pro. The model scored 55 on its Intelligence Index. That score puts it well above the average for similar models. The article also says it is placed on the Pareto line, which means it is one of the best choices for the money compared with other models.

Key numbers

• Intelligence score: 55 (median for comparable models: 27)
• Speed: 77 output tokens per second (average: 81), so it is a bit slower than average
• Price: $2.00 per 1M input tokens and $10.00 per 1M output tokens, which is about the same as the median
• Cost to run the Intelligence Index: $2.11 per task
• Verbosity: 150M output tokens during the test (median: 87M), so it writes a lot

Some background

AI models are computer programs that can read and write text. Companies test them with a set of hard tasks to compare them. Artificial Analysis uses a mix of 10 tests, including coding, science questions, and office-style work, to give each model one score. Kimi K3.1 Pro is a reasoning model, which means it thinks through steps before it answers. It can read text and images, but it only writes text.

Why it matters

This result shows that Kimi K3.1 Pro is among the smartest models that these tests have measured, and that it does not cost much more than others. For users and companies, a model that is smart and reasonably priced can be a good choice. Its slow speed and long answers are downsides to keep in mind.

What to watch

• Whether other tests give the same strong results
• How fast the model becomes in later updates
• Whether prices stay the same for users
• How other AI companies respond with new models at similar prices