MTAC-IFBench - C-ISR (OpenCode): leaderboard
Multi-turn agentic coding instruction following under the OpenCode framework: Constraint-Instruction Success Rate, the share of instructions whose every constraint is satisfied, averaged over turn buckets 1-10. Higher is better. Source: arxiv.org/abs/2609.14992.
Metric: Constraint-Instruction Success Rate (%, all constraints met, average over turn buckets). Source: arxiv.org. Saturation forecast: Around 2029. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GLM-5.2 | 6.7 | #80 |
| 2 | DeepSeek V4 Pro | 5.9 | #112 |
| 3 | GLM-5.1 | 4.5 | #109 |
| 4 | Seed 2.0 Pro | 4.3 | #122 |
| 5 | Qwen 3.6 Plus | 2.4 | #124 |
| 6 | DeepSeek V4 Flash | 1.1 | #155 |
| 7 | Claude Haiku 4.5 | 1 | #287 |
| 8 | Qwen 3.5 27B | 1 | #192 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mtac-ifbench-c-isr-opencode · How It Works · Data refreshed daily, snapshot 2026-11-05.