Classification
Dietary Allergy Menu Screening
Given a patient allergy list, diet order, menu ingredients, and cross-contact rules, classify each meal as allowed, blocked, or requires review.
Healthcare & Life SciencesPer meal
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
275 in · 406 out · 2.51s p50
$10.99 /100K
AWS Bedrock temporary list price
View token rates
AWS Bedrock · $0.0600 input · $0.2300 output /1M
93%
Arcee Trinity Large Thinking
400B · Off-the-shelf
256 in · 760 out · 1.95s p50
$67.20 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $0.2500 input · $0.8000 output /1M
TrustedRouter · $0.2638 input · $0.8440 output /1M
95%
Gemini 3.1 Pro Preview
Hosted · Off-the-shelf
254 in · 293 out · 3.45s p50
$402 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $2.00 input · $12.00 output /1M
TrustedRouter · $2.11 input · $12.66 output /1M
96%
Best accuracy
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 93% | 100/100 | $10.99 |
| Qwen3 235B A22B | mid | 89% | 100/100 | $21.07 |
| DeepSeek V3 | mid | 88% | 100/100 | $10.73 |
| Mistral Large 2407 | mid | 80% | 100/100 | $20.20 |
| Arcee Trinity Large Thinking | mid | 95% | 100/100 | $67.20 |
| GPT-5.4 | frontier | 94% | 100/100 | $126 |
| Claude Opus 4.7 | frontier | 93% | 100/100 | $405 |
| Gemini 3.1 Pro Preview | frontier | 96% | 100/100 | $402 |
| Gemma 4 E4B IT | small | 85% | 100/100 | $0.8208 |
| Granite 4.1 8B | small | 86% | 100/100 | $1.80 |
| Ministral 8B Instruct 2410 | small | 66% | 100/100 | $4.59 |
| Qwen3 4B Instruct 2507 | small | 79% | 100/100 | $9.81 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T07:54:02.740Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock