Classification
Attorney-Client Privilege Triage
Given synthetic business emails, classify likely privilege status, identify attorney involvement and legal-purpose signals, and route uncertain cases for review.
Legal & CompliancePer email
Not benchmarked
No result on this tier
Token profile pending
— /100K
Run this tier to add it
—
Gemma 4 E4B IT
E4B · Off-the-shelf
568 in · 30 out · 0.48s p50
$1.51 /100K
Current price · TrustedRouter
View token rates · observed 9/11/2026
TrustedRouter · $0.0211 input · $0.1055 output /1M
82%
Cheapest
Qwen3 235B A22B
235B (22B active) · Off-the-shelf
556 in · 55 out · 0.98s p50
$35.31 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $0.4550 input · $1.82 output /1M
TrustedRouter · $0.6330 input · $2.11 output /1M
91%
Claude Opus 4.7
Hosted · Off-the-shelf
901 in · 86 out · 2.11s p50
$666 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $5.00 input · $25.00 output /1M
TrustedRouter · $5.28 input · $26.38 output /1M
100%
Best accuracy
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 66% | 100/100 | $20.35 |
| Qwen3 235B A22B | mid | 91% | 100/100 | $35.31 |
| DeepSeek V3 | mid | 89% | 100/100 | $18.83 |
| Mistral Large 2407 | mid | 76% | 100/100 | $41.15 |
| Arcee Trinity Large Thinking | mid | 77% | 100/100 | $102 |
| GPT-5.4 | frontier | 99% | 100/100 | $206 |
| Claude Opus 4.7 | frontier | 100% | 100/100 | $666 |
| Gemini 3.1 Pro Preview | frontier | 97% | 100/100 | $883 |
| Gemma 4 E4B IT | small | 82% | 100/100 | $1.51 |
| Granite 4.1 8B | small | 67% | 100/100 | $3.24 |
| Ministral 8B Instruct 2410 | small | 53% | 100/100 | $9.33 |
| Qwen3 4B Instruct 2507 | small | 77% | 100/100 | $14.96 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-08T20:25:03.151Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock