Published task benchmark

PRD-Draft

Converts a list of features into a formal Product Requirement Document.

Compare measured quality, workload cost, and latency across model tiers. Results are task-specific and directional—not a general model ranking.

EngineeringGenerationEvaluated 9/4/2026100 evaluation examplesFallback pricing checked September 11, 2026
Generation

PRD-Draft

Converts a list of features into a formal Product Requirement Document.

EngineeringPer feature set
Not benchmarked
No result on this tier
Token profile pending
/100K
Run this tier to add it
Ministral 8B Instruct 2410
8B · Off-the-shelf
262 in · 662 out · 16.08s p50
$13.86 /100K
AWS Bedrock temporary proxy price
View token rates
AWS Bedrock · $0.1500 input · $0.1500 output /1M
50%
Mistral Large 2407
123B · Off-the-shelf
288 in · 580 out · 14.13s p50
$107 /100K
Best current price · TrustedRouter
Compare 2 router rates · observed 9/11/2026
TrustedRouter · $0.5275 input · $1.58 output /1M
OpenRouter · $2.00 input · $6.00 output /1M
91%
Gemini 3.1 Pro Preview
Hosted · Off-the-shelf
256 in · 1,314 out · 11.32s p50
$1628 /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
99%
Best accuracy
All 12 model results and methodology

Generated 100/100 examples

ModelTierQualityJudgedScenario cost
Nemotron Nano 9B v2small4%100/100$39.85
Qwen3 235B A22Bmid0%100/100$216
DeepSeek V3mid1%100/100$82.63
Mistral Large 2407mid91%100/100$107
Arcee Trinity Large Thinkingmid0%100/100$132
GPT-5.4frontier97%100/100$1607
Claude Opus 4.7frontier75%100/100$3276
Gemini 3.1 Pro Previewfrontier99%100/100$1628
Gemma 4 E4B ITsmall22%100/100$6.95
Granite 4.1 8Bsmall23%100/100$8.08
Ministral 8B Instruct 2410small50%100/100$13.86
Qwen3 4B Instruct 2507small0%100/100$73.16

LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-04T21:45:03.208Z.

Directional: measured on a synthetic eval set generated by drydock. Cost/latency are not yet captured for taskrouter-run benchmarks.

drydock

Have a neighboring workload?

Describe it in Task Explorer. If it is not measured yet, you can request a benchmark and follow future results.

Explore another task