Measurable IEP Goal Drafting
Given anonymized baseline assessment data and a target skill, draft measurable goals with condition, behavior, criterion, and review period for specialist review.
View token rates · observed 9/11/2026
Compare 2 router rates · observed 9/11/2026
Compare 2 router rates · observed 9/11/2026
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 10% | 100/100 | $26.00 |
| Qwen3 235B A22B | mid | 3% | 100/100 | $87.95 |
| DeepSeek V3 | mid | 35% | 100/100 | $28.99 |
| Mistral Large 2407 | mid | 14% | 100/100 | $79.60 |
| Arcee Trinity Large Thinking | mid | 23% | 100/100 | $64.03 |
| GPT-5.4 | frontier | 54% | 100/100 | $607 |
| Claude Opus 4.7 | frontier | 36% | 100/100 | $2741 |
| Gemini 3.1 Pro Preview | frontier | 45% | 100/100 | $2018 |
| Gemma 4 E4B IT | small | 38% | 100/100 | $2.34 |
| Granite 4.1 8B | small | 3% | 100/100 | $4.36 |
| Ministral 8B Instruct 2410 | small | 11% | 100/100 | $9.09 |
| Qwen3 4B Instruct 2507 | small | 4% | 100/100 | $27.39 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T09:33:02.922Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock