Extraction
Grounded Contract Question Answering
Answer questions only from supplied authoritative contract chunks, cite the supporting chunk, and abstain when the answer is absent or superseded. Available model: neurometric/grounded-document-qa.
Legal & CompliancePer contract
Not benchmarked
No result on this tier
Token profile pending
— /100K
Run this tier to add it
—
Granite 4.1 8B
8B · Off-the-shelf
406 in · 180 out · 2.19s p50
$3.83 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/4/2026
OpenRouter · $0.0500 input · $0.1000 output /1M
TrustedRouter · $0.0527 input · $0.1055 output /1M
27%
Mistral Large 2407
123B · Off-the-shelf
459 in · 200 out · 5.84s p50
$55.86 /100K
Best current price · TrustedRouter
Compare 2 router rates · observed 9/11/2026
TrustedRouter · $0.5275 input · $1.58 output /1M
OpenRouter · $2.00 input · $6.00 output /1M
63%
GPT-5.4
Hosted · Off-the-shelf
400 in · 164 out · 2.13s p50
$346 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $2.50 input · $15.00 output /1M
TrustedRouter · $2.64 input · $15.82 output /1M
84%
Best accuracy
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 13% | 100/100 | $13.96 |
| Qwen3 235B A22B | mid | 59% | 100/100 | $48.41 |
| DeepSeek V3 | mid | 53% | 100/100 | $19.58 |
| Mistral Large 2407 | mid | 63% | 100/100 | $55.86 |
| Arcee Trinity Large Thinking | mid | 39% | 100/100 | $109 |
| GPT-5.4 | frontier | 84% | 100/100 | $346 |
| Claude Opus 4.7 | frontier | 82% | 100/100 | $1582 |
| Gemini 3.1 Pro Preview | frontier | 59% | 100/100 | $798 |
| Gemma 4 E4B IT | small | 17% | 100/100 | $1.79 |
| Granite 4.1 8B | small | 27% | 100/100 | $3.83 |
| Ministral 8B Instruct 2410 | small | 14% | 100/100 | $7.72 |
| Qwen3 4B Instruct 2507 | small | 14% | 100/100 | $10.07 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-08T20:45:02.796Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock