Benchmark profile
Kimi Code Bench v2
A Moonshot AI internal coding-agent benchmark for realistic software-engineering tasks across mainstream programming languages and production technology stacks.
Data verifiedBenchmark score on Kimi Code Bench v2 — July 27, 2026
BenchLM mirrors the published score view for Kimi Code Bench v2. Kimi K3 leads the public snapshot at 72.9% , followed by Kimi K2.7 Code (62.0%). BenchLM does not use these results to rank models overall.
Kimi K3
Moonshot AI
kimi-3
Kimi K2.7 Code
Moonshot AI
kimi-k2-7-code
Benchmark score table (2 models)
ScoreAbout Kimi Code Bench v2
Year
2026
Tasks
Realistic coding-agent tasks
Format
Coding-agent pass rate
Difficulty
Production software engineering
Moonshot describes Kimi Code Bench v2 as an in-house coding-agent benchmark covering backend services, infrastructure, performance engineering, systems programming, security, frontend development, and ML/data engineering. BenchLM stores provider-reported exact values as display-only launch evidence.
BenchLM freshness & provenance
Version
Kimi Code Bench v2 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
FAQ
What does Kimi Code Bench v2 measure?
A Moonshot AI internal coding-agent benchmark for realistic software-engineering tasks across mainstream programming languages and production technology stacks.
Which model scores highest on Kimi Code Bench v2?
Kimi K3 by Moonshot AI currently leads with a score of 72.9% on Kimi Code Bench v2.
How many models are evaluated on Kimi Code Bench v2?
2 AI models have been evaluated on Kimi Code Bench v2 on BenchLM.
Compare Top Models on Kimi Code Bench v2
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
One email each week. Unsubscribe anytime.