Model Profile
Kimi K2 Thinking
Use this page to decide where this model is a strong fit. Rankings below are benchmark-backed by use case, with explicit confidence and contributor metrics.
Identity
ID: external/kimi/kimi-k2-thinking
Author: kimi
Origin: external_benchmark_shadow
Arch: unknown
Benchmark Coverage
Scored use cases: 12
Avg confidence: 34.9%
Evidence points: 242
Raw rows: 535
Weighted rows: 39
Catalog Metadata
Parameters: unknown
Context window: 4096
Downloads: 0
Price / 1M tokens: $1.07 (blended 3:1)
Intelligence Profile
Dimension Breakdown
No eq benchmarks found
* Low confidence — limited benchmark evidence for this dimension
4/5 dimensions scored · Last updated Apr 21, 2026
Benchmark Signals
Click through to the benchmark source behind this model profile.
SWE-bench Verified Leaderboard
swe_verified_resolved_pct
Normalized value 79.7% · confidence 100.0%
Strongest impact in CAD scripting helper
swebench_verified_official.swe_verified_resolved_pct · Apr 1, 2026
Sonar Java Quality Leaderboard
functional_skill_pct
Normalized value 85.8% · confidence 100.0%
Strongest impact in Code Review Assistant
sonar_java_quality.functional_skill_pct · Apr 1, 2026
Vals CorpFin v2
overall_accuracy_pct
Normalized value 77.7% · confidence 100.0%
Strongest impact in Thesis red teaming
vals_corp_fin_v2.overall_accuracy_pct · Mar 31, 2026
Vals Finance Agent
overall_accuracy_pct
Normalized value 57.6% · confidence 100.0%
Strongest impact in Thesis red teaming
vals_finance_agent.overall_accuracy_pct · Mar 31, 2026
Vals Tax Eval v2
overall_accuracy_pct
Normalized value 84.7% · confidence 100.0%
Strongest impact in Accounts payable invoice extraction (text)
vals_tax_eval_v2.overall_accuracy_pct · Mar 31, 2026
Vals CorpFin v2
shared_max_context_accuracy_pct
Normalized value 82.5% · confidence 100.0%
Strongest impact in Thesis red teaming
vals_corp_fin_v2.shared_max_context_accuracy_pct · Mar 31, 2026
Coverage Diagnostics
actively scoredUse-Case Scores
138
Total Measurements
535
Weighted Measurements
39
Weighted Sources
21
Raw Source Coverage
Weighted Source Coverage
Best Use Cases for This Model
| Use Case | Score |
|---|---|
| Verilog/VHDL generation use_case.eda.verilog_generation | 25.4% |
| Refactoring assistant use_case.dev.refactoring | 24.5% |
| Quant research code generation use_case.fin.alpha_research_codegen | 24.1% |
| Debugging assistant use_case.dev.debugging | 23.3% |
| Code Review Assistant use_case.dev.code_review_assistant | 21.2% |
| Thesis red teaming use_case.fin.thesis_red_team | 20.3% |
| CAD scripting helper use_case.eng.cad_scripting_helper | 19.9% |
| Integration test generation use_case.dev.integration_tests | 19.8% |
| Simulation setup assistant use_case.eng.simulation_setup_assistant | 19.2% |
| Code generation use_case.dev.code_generation | 19.0% |
| Accounts payable invoice extraction (text) use_case.fin.ap_invoice_extraction | 18.8% |
| Earnings call synthesis use_case.fin.earnings_call_synthesis | 18.3% |