Models / Moonshot AI
Kimi K3
by Moonshot AI ·
moonshotai/kimi-k3Kimi K3 is a 2.8T parameter open-weight multimodal reasoning model from Moonshot AI. It is suited for complex coding, knowledge work, and long-horizon agentic workflows, and is particularly strong at...
reasoningtool-usevisionreleased 2026-07-16
Legal score · raw
-
not yet benchmarked
Context
1.05M
max output 944K
Input
$3
per 1M tokens
Output
$15
per 1M tokens
Suite cost
-
run the suite to see
Benchmark results
| Task | Category | Raw | Juiced | Correct | Latency | Cost | Ran |
|---|---|---|---|---|---|---|---|
| Hearsay Identification | Evidence | - | - | - | - | - | - |
| Bluebook Citation Format | Research & Writing | - | - | - | - | - | - |
| Federal Civil Procedure | Procedure | - | - | - | - | - | - |
| Limitations Arithmetic | Procedure | - | - | - | - | - | - |
| Contract Clause Classification | Contracts | - | - | - | - | - | - |
| Citation Hallucination Resistance | Reliability | - | - | - | - | - | - |
Measured by DocketBuster
These numbers come from DocketBuster's own legal battery, not from DocketRouter's suite. Latest run per metric, with n and a 95% Wilson interval where the source reports one. See docketbuster.com/benchmarks.
| Metric | Value | n | Interval | Measured |
|---|---|---|---|---|
| Statute pinpoint, exact section (no retrieval) | 25.0% (75/300) | 300 | 95% CI 20.4% to 30.2% | 2026-08-22 |
| Say-nothing rate (declines to bluff when the answer is not in the record) | 100.0% | 58 | 95% CI 93.8% to 100.0% | 2026-08-22 |
| Abstained on statute pinpoint | 47.7% (143/300) | 300 | count, no interval reported | 2026-08-22 |
| Coaching quality (GW-14x, 0 to 8) | 7.36 / 8 | - | rubric mean, no interval reported | 2026-08-23 |
Source files: hard-llm-ortier-kimik3.json, gw14x-ortier-kimik3.json. Raw model name in source: moonshotai/kimi-k3.