AI Benchmark Model

Kimi K3 (low)

Benchmark scores, pricing, speed, and model comparisons for Kimi K3 (low).

Key scores

Intelligence46.6Index score
Coding72.0Index score
Speed33.7Tokens/sec
Blended Cost$6.00Per 1M tokens

The Kimi K3 (low) model, developed by Kimi and released on July 16, 2026, is a language model designed for specific computational and coding tasks. In performance evaluations, the model achieved an intelligence index of 46.6 and a coding index of 72. Benchmark results for the model include a GPQA score of 0.842, an HLE score of 0.24, a SciCode score of 0.512, and an LCR score of 0.693333333333333. Regarding operational efficiency, the model demonstrates an output speed of 33.686 tokens per second with a time to first token of 3.352 seconds. Pricing for the Kimi K3 (low) is structured at $3.00 per 1 million input tokens and $15.00 per 1 million output tokens, resulting in a blended rate of $6.00 per 1 million tokens.

Benchmark results

4 tracked benchmark rows for Kimi K3 (low).

GPQA Reasoning
84.2%
LCR Other
69.3%
SciCode Other
51.2%
HLE Other
24.0%

Comments (0)

No comments yet

Be the first to share your thoughts!