Generation
Negative Product Review Response
Given a customer review and response policy, draft an empathetic reply that acknowledges the issue, avoids prohibited admissions, and offers an approved next step.
Sales & MarketingPer review
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
345 in · 123 out · 1.91s p50
$2.03 /100K
Current price · TrustedRouter
View token rates · observed 9/11/2026
TrustedRouter · $0.0211 input · $0.1055 output /1M
96%
Cheapest
DeepSeek V3
Hosted · Off-the-shelf
335 in · 126 out · 1.80s p50
$23.14 /100K
Best current price · TrustedRouter
Compare 2 router rates · observed 9/11/2026
TrustedRouter · $0.3376 input · $0.9390 output /1M
OpenRouter · $0.2900 input · $1.14 output /1M
99%
GPT-5.4
Hosted · Off-the-shelf
329 in · 110 out · 1.94s p50
$247 /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
100%
Best accuracy
All 12 model results and methodology
Generated 100/100 examples
| Model | Tier | Quality | Judged | Scenario cost |
|---|---|---|---|---|
| Nemotron Nano 9B v2 | small | 40% | 100/100 | $16.24 |
| Qwen3 235B A22B | mid | 97% | 100/100 | $43.91 |
| DeepSeek V3 | mid | 99% | 100/100 | $23.14 |
| Mistral Large 2407 | mid | 94% | 100/100 | $46.74 |
| Arcee Trinity Large Thinking | mid | 92% | 100/100 | $37.29 |
| GPT-5.4 | frontier | 100% | 100/100 | $247 |
| Claude Opus 4.7 | frontier | 100% | 100/100 | $909 |
| Gemini 3.1 Pro Preview | frontier | 98% | 100/100 | $1345 |
| Gemma 4 E4B IT | small | 96% | 100/100 | $2.03 |
| Granite 4.1 8B | small | 83% | 100/100 | $3.71 |
| Ministral 8B Instruct 2410 | small | 77% | 100/100 | $7.74 |
| Qwen3 4B Instruct 2507 | small | 84% | 100/100 | $15.09 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T03:24:02.845Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock