Best AI model for Software & DevOps
DeepSeek V4.1 Flash (Venice) ranks first for Software & DevOps with a score of 89.1, passing 100% of its 7 runs.
Lowest average cost per success: Qwen2.5 7B (local) at $0 (local).
Based on 106 logged runs across 14 use cases; 10 models had enough runs to rank.
| # | Model | Score | Runs | Pass rate | Cost / success | Avg time | Step-ins / run |
|---|---|---|---|---|---|---|---|
| 1 | DeepSeek V4.1 Flash (Venice) | 89.1 | 7 | 100% (7/7) | $0.0078 | 2m | 0.00 |
| 2 | Claude Haiku 5.5 | 89.1 | 18 | 94% (17/18) | $0.02 | 46s | 0.11 |
| 3 | Gemma 4 31B Instruct (Venice) | 88.1 | 7 | 100% (7/7) | $0.0033 | 1m | 0.00 |
| 4 | Claude Sonnet 5.5 | 84.7 | 23 | 100% (23/23) | $0.31 | 4m | 0.13 |
| 5 | Grok 4.3 (Venice) | 83.8 | 7 | 86% (6/7) | $0.01 | 15s | 0.00 |
| 6 | GLM 5.3 (Venice) | 78.3 | 5 | 100% (5/5) | $0.04 | 1m | 0.00 |
| 7 | Qwen3 Coder 480B Turbo (Venice) | 76.7 | 7 | 71% (5/7) | $0.0051 | 23s | 0.00 |
| 8 | Claude Opus 5.5 | 75.4 | 21 | 100% (21/21) | $0.66 | 24m | 0.14 |
| 9 | Qwen2.5 7B (local) | 49.4 | 4 | 25% (1/4) | $0 (local) | 33s | 0.00 |
| 10 | Llama 3.2 3B (local) | 24.4 | 4 | 0% (0/4) | - | 11s | 0.00 |
Only models with 3+ runs here are ranked; all settings of a model are grouped. Scores use the default weights (methodology).
Browse software & devops use cases → · Filtered leaderboard →