Kimi K3 Tech Report - DECK Bench: leaderboard
Moonshot AI's in-house DECK Bench measures producing high-quality presentation decks from task descriptions drawn from real usage scenarios. Kimi K3 technical report, Table 3; score in percent, higher is better; models at maximum reasoning effort (GPT-5.5 at xhigh). The harness column reads N/A.
Metric: Score (%). Source: raw.githubusercontent.com. Saturation forecast: Around July 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol (Max) | 74.7 | #22 (GPT-5.6 Sol) |
| 2 | Kimi K3 (Max) | 73.5 | #35 (Kimi K3) |
| 3 | Claude Fable 5 (Max) | 73 | #20 (Claude Fable 5) |
| 4 | GLM-5.2 (Max) | 68.6 | #80 (GLM-5.2) |
| 5 | GPT-5.5 (xHigh) | 68.2 | #33 (GPT-5.5) |
| 6 | Claude Opus 4.8 (Max) | 66.9 | #39 (Claude Opus 4.8) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=kimi-k3-tech-report-deck-bench · How It Works · Data refreshed daily, snapshot 2026-11-05.