PLPramaana

Moonshot AI

Kimi K2.7 Code

Release date: 6/12/2026

A long-context coding specialist optimized for repositories and autonomous editing.

Pramaana IndexAccuracy: StandardLatency: StrongCost: Leading
Accuracy / Pramaana Index
62.34%± 0.99
Latency / Pramaana Index
314.28s
Cost / Test / Pramaana Index
$0.29
Context window
256k
Max output tokens
64k
Input modality
Text · File
Hyperparameter settings

temperature: 0.2

reasoning: extended

fallbacks: disabled

Benchmark performance for Kimi K2.7 Code

BenchmarksAccuracyRanking
Pramaana Index62.34% ± 0.9912 / 35
Pramaana Multimodal Index58.26% ± 1.1612 / 24
Legal Research Fidelity56.74% ± 1.3312 / 18
Financial Modeling Agent52.45% ± 1.5012 / 22
Clinical Documentation Safety56.04% ± 1.6712 / 19
MATH-50070.14% ± 0.4812 / 112
GPQA Diamond74.15% ± 0.6512 / 121
MMLU-Pro71.64% ± 0.8212 / 119
Adaptive Tutoring Dialogue57.94% ± 0.9912 / 16
SWE-bench Verified56.05% ± 1.166 / 69
LiveCodeBench72.96% ± 1.337 / 127
Agentic Reliability46.46% ± 1.5012 / 12
Strategic Games Arena54.55% ± 1.6712 / 20
Benefits Navigator57.85% ± 0.4812 / 16

License: Modified MIT · Evaluation settings, model endpoints, and benchmark versions are frozen with each published run.

Read our methodology