Classification
Return Abuse Risk Triage
Given anonymized order and return history plus policy thresholds, classify a return as standard, review, or high risk and cite the triggering signals.
E-commerce & LogisticsPer return
Not benchmarked
No result on this tier
Token profile pending
— /100K
Run this tier to add it
—
Nemotron Nano 9B v2
9B · Off-the-shelf
570 in · 714 out · 4.46s p50
$19.84 /100K
AWS Bedrock temporary list price
View token rates
AWS Bedrock · $0.0600 input · $0.2300 output /1M
86%
Qwen3 235B A22B
235B (22B active) · Off-the-shelf
543 in · 47 out · 0.70s p50
$33.26 /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
83%
GPT-5.4
Hosted · Off-the-shelf
502 in · 62 out · 1.55s p50
$219 /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
92%
Best accuracy
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 86% | 100/100 | $19.84 |
| Qwen3 235B A22B | mid | 83% | 100/100 | $33.26 |
| DeepSeek V3 | mid | 76% | 100/100 | $18.96 |
| Mistral Large 2407 | mid | 77% | 100/100 | $43.10 |
| Arcee Trinity Large Thinking | mid | 65% | 100/100 | $135 |
| GPT-5.4 | frontier | 92% | 100/100 | $219 |
| Claude Opus 4.7 | frontier | 89% | 100/100 | $593 |
| Gemini 3.1 Pro Preview | frontier | 88% | 100/100 | $1104 |
| Gemma 4 E4B IT | small | 65% | 100/100 | $1.60 |
| Granite 4.1 8B | small | 49% | 100/100 | $3.21 |
| Ministral 8B Instruct 2410 | small | 37% | 100/100 | $9.27 |
| Qwen3 4B Instruct 2507 | small | 47% | 100/100 | $16.42 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T10:16:02.553Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock