Multi-Challenge
Multi-Challenge - multi-turn instruction following benchmark by ScaleAI.
Category
Instruction Following
Max Score
100.0
Score Type
percent
Active
Yes
Model Scores
Scores for Multi-Challenge
| Model | Score | Date | Verified |
|---|---|---|---|
| NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 |
98.61%
|
04.06.2026 | Verified |
| MiniMax-M2.7 |
42.50
|
04.06.2026 | Unverified |
| GLM-5.1 |
94.91%
|
04.06.2026 | Unverified |
| Kimi K2.6 |
95.37%
|
04.06.2026 | Unverified |
| Qwen3.5-397B-A17B |
99.07%
|
04.06.2026 | Unverified |
| DeepSeek-V4-Pro |
100.00%
|
04.06.2026 | Unverified |
| DeepSeek-V4-Flash |
97.22%
|
04.06.2026 | Unverified |
| Qwen3.5-4B |
30.09%
|
01.02.2026 | Verified |
| Qwen3.5-122B-A10B |
87.96%
|
01.02.2026 | Verified |
| Qwen3.5-35B-A3B |
81.02%
|
01.02.2026 | Unverified |
| Qwen3.5-27B |
84.72%
|
01.02.2026 | Unverified |
| Qwen3-235B-A22B |
35.65%
|
01.02.2026 | Unverified |
| GPT-5-mini 2025-08-07 |
76.39%
|
01.02.2026 | Unverified |