MMoonshot | Kimi K3
Kimi 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 navigating large repositories, using tools, debugging, and iterating against images, logs, tests, and runtime feedback. Its architecture uses KDA and Attention Residuals for computational efficiency.
Fair
Significant guardrail issues may impede work.
Strengths
- Excellent at 2.1. Exclusivity of Jesus
- Excellent at 3.3. The Crucifixion
Weaknesses
- Limited task completion capability
- Struggles with 1.1. Missiological Research
- Struggles with 1.6. Problematic Vocabulary
- Struggles with 1.7. Difficult Passages
- Struggles with 2.3. Reality of Judgment
Kimi 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 navigating large repositories, using tools, debugging, and iterating against images, logs, tests, and runtime feedback. Its architecture uses KDA and Attention Residuals for computational efficiency.
Moonshot
moonshotai/kimi-k3
1
Categories
Recent Tests
| Date | Score | Version | Tier 1 (Task) | Tier 2 (Gospel) | Tier 3 (Worldview) | Trust Tier |
|---|---|---|---|---|---|---|
| 7/17/2026 | 46.0 | 1.0.0 | 42.9 | 51.7 | 56.7 | automated |