Agents · Benchmark profile
Artificial Analysis private agentic knowledge-work evaluation with Elo over rubric pass rate, analytical quality, and presentation quality on business deliverables.
Data verified 16 Jul 2026 · Methodology 1.8.0
Benchmark score on AA-Briefcase
Kimi K3 leads at 152.7 elo-proxy, followed by Gemini 3.6 Flash (96.1 elo-proxy) and Muse Spark 1.1 (86.3 elo-proxy).
4 ranked models · higher is better · labels show rank and score
| Model | Rank | Provider | Score |
|---|---|---|---|
| Kimi K3 | #1 | Moonshot AI | 152.7 elo-proxy |
| Gemini 3.6 Flash | #2 | 96.1 elo-proxy | |
| Muse Spark 1.1 | #3 | Meta | 86.3 elo-proxy |
| Gemini 3.5 Flash-Lite | #4 | 63.4 elo-proxy |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Kimi K3 kimi-k3 | Moonshot AI | closed | Reference only | 152.7 elo-proxy |
| #2 | Gemini 3.6 Flash gemini-3.6-flash | closed | Reference only | 96.1 elo-proxy | |
| #3 | Muse Spark 1.1 muse-spark-1-1 | Meta | closed | Reference only | 86.3 elo-proxy |
| #4 | Gemini 3.5 Flash-Lite gemini-3.5-flash-lite | closed | Reference only | 63.4 elo-proxy |
The top of this snapshot is led by Kimi K3 at 152.7 elo-proxy; third place is 66.4 points behind. The top-4 spread is 89.3 points.
About AA-Briefcase
Artificial Analysis private agentic knowledge-work evaluation with Elo over rubric pass rate, analytical quality, and presentation quality on business deliverables. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
Artificial Analysis private agentic knowledge-work evaluation with Elo over rubric pass rate, analytical quality, and presentation quality on business deliverables.
Kimi K3 by Moonshot AI currently leads with 152.7 elo-proxy.
4 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
