Kimi K3.1 Pro Scores 55 on Artificial Analysis Intelligence Index

• Kimi K3.1 Pro achieves an Artificial Analysis Intelligence Index score of 55, placing it 4th among 215 models and significantly above the median of 27 for similar price tiers.
• The model is positioned on the Pareto line in the Intelligence vs. Cost chart, offering a competitive cost of $2.00 per 1M input tokens and $10.00 per 1M output tokens.
• Performance is characterized by a slow output speed of 57.5 tokens/sec and high verbosity, generating 150M output tokens during evaluation compared to a median of 87M.
• Technical specifications include a 1.0M token context window, support for text and image inputs, and a proprietary 2.8 trillion parameter architecture released in October 2026.

Performance Metrics

• Time To First Token (TTFT) is 3.07 seconds, faster than the 3.90 second median for comparable models.
• The blended rate for cache hit, input, and output (7:2:1 ratio) is $1.54 per 1M tokens.
• Output speed is measured after the first chunk is received, excluding Time To First Token and overhead.

Evaluation Methodology

• The Intelligence Index v4.3.2 aggregates results from 10 benchmarks, including AA-Briefcase v1.1, GDPval-AA v2.1, and Terminal-Bench 4.0.
• Additional metrics include the AA-Omniscience Index for knowledge reliability and AA-PersonalityBench for personality traits.