Summarization
Customer-Facing Release Note Synthesis
Given issue tickets and merged pull-request summaries, produce release notes grouped by feature, fix, and breaking change while excluding internal-only work.
Engineering & ProductPer release
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
717 in · 926 out · 6.38s p50
$25.60 /100K
AWS Bedrock temporary list price
View token rates
AWS Bedrock · $0.0600 input · $0.2300 output /1M
92%
Arcee Trinity Large Thinking
400B · Off-the-shelf
640 in · 747 out · 2.35s p50
$75.76 /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
92%
GPT-5.4
Hosted · Off-the-shelf
657 in · 229 out · 2.53s p50
$508 /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 | 92% | 100/100 | $25.60 |
| Qwen3 235B A22B | mid | 6% | 100/100 | $112 |
| DeepSeek V3 | mid | 85% | 100/100 | $43.16 |
| Mistral Large 2407 | mid | 75% | 100/100 | $93.58 |
| Arcee Trinity Large Thinking | mid | 92% | 100/100 | $75.76 |
| GPT-5.4 | frontier | 100% | 100/100 | $508 |
| Claude Opus 4.7 | frontier | 90% | 100/100 | $1861 |
| Gemini 3.1 Pro Preview | frontier | 99% | 100/100 | $1583 |
| Gemma 4 E4B IT | small | 86% | 100/100 | $3.63 |
| Granite 4.1 8B | small | 12% | 100/100 | $6.79 |
| Ministral 8B Instruct 2410 | small | 3% | 100/100 | $15.53 |
| Qwen3 4B Instruct 2507 | small | 9% | 100/100 | $35.55 |
LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-09T01:09:03.056Z.
Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.
drydock