Thinking Machines Inkling-Small Launch - StrongREJECT: leaderboard
StrongREJECT unambiguously harmful requests run by Thinking Machines: the rate at which a model refuses them. Thinking Machines' Inkling-Small launch post (July 2026), Safety rows of the full comparison table; percent, higher is better; Inkling models at effort 0.99. Scores cluster between 97% and 100%, so differences are small.
Metric: Refusal Rate (%). Source: thinkingmachines.ai. Saturation forecast: Estimated already saturated. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 397B A17B | 99.4 | #154 |
| 2 | MiniMax-M2.7 | 99.4 | #270 |
| 3 | MiMo-V2.5 | 99.3 | #196 |
| 4 | GPT-5.6 Luna | 98.7 | #79 |
| 5 | Nemotron 3 Ultra | 98.7 | #245 |
| 6 | Claude Haiku 4.5 | 98.6 | #287 |
| 7 | Gemini 3.5 Flash Lite | 97.6 | #174 |
| 8 | DeepSeek V4 Flash | 97.4 | #155 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=thinking-machines-inkling-small-launch-strongreject · How It Works · Data refreshed daily, snapshot 2026-11-05.