MT-Bench
AI benchmark results and model performance
Type: numericalCommunication
Recorded results
17
Average score
2.12
Score range
0.09 - 8.61
Leading model
8.61 - Granite 4.1 30B
Scores over time
Individual benchmark scores plotted by date.
Model results
| Organisation | Models | Reported | Top score | Info | Self reported | Source |
|---|---|---|---|---|---|---|
| Granite 4.1 30B | 29 Apr 2026 | 8.61 | Avg | Yes | Source | |
| Granite 4.1 8B | 29 Apr 2026 | 8.61 | Avg | Yes | Source | |
| Granite 4.1 3B | 29 Apr 2026 | 7.57 | Avg | Yes | Source | |
| Qwen 72B | - | 0.94 | inferred family alias from qwen-2.5-72b-instruct (score=0.3060; benches=14) | Yes | Source | |
| Llama 3.3 Nemotron Super 49B v1 | 18 Mar 2025 | 0.92 | - | Yes | Source | |
| Llama 3.3 Nemotron Super 49B V1.5 | - | 0.92 | inferred version-family alias from llama-3.3-nemotron-super-49b-v1 | Yes | Source | |
| DeepSeek V2.5 (2024-09-05) | 05 Sept 2024 | 0.90 | inferred alias from deepseek-v2.5 | Yes | Source | |
| DeepSeek V2.5 (2024-12-10) | 10 Dec 2024 | 0.90 | inferred alias from deepseek-v2.5 | Yes | Source | |
| Qwen 7B | - | 0.88 | inferred family alias from qwen-2.5-7b-instruct (score=0.3083; benches=14) | Yes | Source | |
| Mistral Large 2.1 | 18 Nov 2024 | 0.86 | inferred version-family alias from mistral-large-2-2407 | Yes | Source | |
| Mistral OCR 2 | 22 May 2025 | 0.86 | inferred family alias from mistral-large-2-2407 (score=0.3100; benches=5) | Yes | Source | |
| Qwen 2 Audio 7B | - | 0.84 | inferred modality/version alias from qwen2-7b-instruct | Yes | Source | |
| Qwen 2 Math 7B | - | 0.84 | inferred high-confidence family alias from qwen2-7b-instruct (score=0.4706; benches=14) | Yes | Source | |
| Llama 3.1 Nemotron Nano 4B V1.1 | - | 0.81 | inferred high-confidence family alias from llama-3.1-nemotron-nano-8b-v1 (score=0.5523; benches=7) | Yes | Source | |
| Llama 3.1 Nemotron Nano 8B V1 | 18 Mar 2025 | 0.81 | - | Yes | Source | |
| Pixtral 12B | 17 Sept 2024 | 0.77 | inferred version-family alias from pixtral-12b-2409 | Yes | Source | |
| Llama 3.1 Nemotron 70B Instruct | 01 Oct 2024 | 0.09 | - | Yes | Source |
