Rankings / Unknown
Kimi K2 Thinking
released 2025-11-06
BenchAtlas Index
as of 2026-07-14
base configuration
rank #139
10 families · 5 categories · high
Benchmark evidence
11 results
agentic coding
| τ²-Bench | 93.0 % | independent | Artificial Analysis |
coding
| LiveCodeBench v6 implementation=artificial-analysis | 85.3 % | independent | Artificial Analysis | |
| SciCode | 42.4 % | independent | Artificial Analysis |
external indices
| Intelligence Index v4.1 | 32.7 points | independent | Artificial Analysis | |
| AA Math Index | 94.7 points | independent | Artificial Analysis |
knowledge science
| GPQA Diamond implementation=artificial-analysis | 83.8 % | independent | Artificial Analysis | |
| Humanity's Last Exam implementation=artificial-analysis | 22.3 % | independent | Artificial Analysis | |
| MMLU-Pro implementation=artificial-analysis | 84.8 % | independent | Artificial Analysis |
long context instruction
| AA-LCR | 66.3 % | independent | Artificial Analysis | |
| IFBench | 68.1 % | independent | Artificial Analysis |
reasoning math
| AIME year=2025 · implementation=artificial-analysis | 94.7 % | independent | Artificial Analysis |
Agent + model results
systems, not bare-model scores
| agent + model mini-SWE-agent + Kimi K2 Thinking | SWE-bench bash-only | 63.4 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Kimi K2 Thinking | SWE-bench Verified | 63.4 % | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Kimi K2 Thinking | SWE-bench bash-only | 0.4 usd_per_task | community | SWE-bench Leaderboard |
| agent + model mini-SWE-agent + Kimi K2 Thinking | SWE-bench Verified | 0.4 usd_per_task | community | SWE-bench Leaderboard |
| agent + model Artificial Analysis harness + Kimi K2 Thinking | Terminal-Bench Hard | 31.1 % | independent | Artificial Analysis |
These scores measure the whole agent system (scaffold, tools, budgets) — they are never merged into the bare model’s numbers.