Extraction
Natural-Language Product Filter Extraction
Convert a shopper query into only allowed search filters such as color, fit, material, price range, and occasion while preserving explicit preferences.
E-commerce & LogisticsPer query
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
255 in · 58 out · 0.92s p50
$1.15 /100K
Current price · TrustedRouter
View token rates · observed 9/11/2026
TrustedRouter · $0.0211 input · $0.1055 output /1M
76%
Cheapest
DeepSeek V3
Hosted · Off-the-shelf
239 in · 46 out · 0.83s p50
$12.18 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $0.2900 input · $1.14 output /1M
TrustedRouter · $0.3376 input · $0.9390 output /1M
90%
Best accuracy
Claude Opus 4.7
Hosted · Off-the-shelf
373 in · 64 out · 1.62s p50
$347 /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
87%
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 47% | 100/100 | $12.28 |
| Qwen3 235B A22B | mid | 76% | 100/100 | $20.16 |
| DeepSeek V3 | mid | 90% | 100/100 | $12.18 |
| Mistral Large 2407 | mid | 64% | 100/100 | $31.81 |
| Arcee Trinity Large Thinking | mid | 66% | 100/100 | $93.90 |
| GPT-5.4 | frontier | 86% | 100/100 | $129 |
| Claude Opus 4.7 | frontier | 87% | 100/100 | $347 |
| Gemini 3.1 Pro Preview | frontier | 67% | 100/100 | $962 |
| Gemma 4 E4B IT | small | 76% | 100/100 | $1.15 |
| Granite 4.1 8B | small | 74% | 100/100 | $1.64 |
| Ministral 8B Instruct 2410 | small | 63% | 100/100 | $4.54 |
| Qwen3 4B Instruct 2507 | small | 66% | 100/100 | $6.78 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T10:07:03.061Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock