BLXBench Speed category
Code · 2026-04-25
Speed category component of the BLXBench public leaderboard. The source describes this category as throughput and TTFT-focused generation tasks; scores are category score percentages.
Top models (higher is better)
| Model | Score |
|---|---|
| Grok 4.3 | 97.8 |
| Gemini 3.1 Flash-Lite | 97.2 |
| Granite 4.1 8B | 97.2 |
| Grok Build 0.1 | 97.2 |
| DeepSeek-V4-Flash | 96.1 |
| Qwen3.7-Plus | 96.1 |
| Mistral Small 4 | 95.6 |
| MiMo-V2.5 | 94.4 |
| Nemotron 3 Nano 30B A3B | 94.4 |
| GLM-5.1 | 93.9 |
| Kimi K2.6 | 93.9 |
| GLM-5.2 | 92.2 |
| Mistral Medium 3.5 | 92.2 |
| Qwen3.6-Flash-2026-04-16 | 92.2 |
| DeepSeek-V4-Pro | 91.7 |
| Nemotron 3 Super 120B A12B | 91.1 |
| North Mini Code | 91.1 |
| Qwen3.7-Max | 91.1 |
| GPT-5.3-Codex | 89.4 |
| GPT-5.5 | 89.4 |
| MiMo-V2.5-Pro | 89.4 |
| CoBuddy | 88.3 |
| Opus 4.7 | 87.8 |
| Kimi K2.7 Code | 87.2 |
| Nemotron 3 Nano Omni 30B A3B | 87.2 |
| MiniMax M3 | 86.7 |
| Claude Fable 5 | 85.6 |
| Opus 4.8 | 84.4 |
| Ring 2.6 1T | 84.2 |
| MiniMax M2.7 | 59.4 |
| Step 3.7 Flash | 52.8 |
| Gemini 3.5 Flash | 46.1 |