Published task benchmark

Financial Variance Investigation Agent

Query supporting records, explain material variances, and open follow-up items with evidence.

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

Finance, accounting, and procurement operationsAnalysisEvaluated 9/11/2026100 evaluation examplesFallback pricing checked September 11, 2026
Analysis

Financial Variance Investigation Agent

Query supporting records, explain material variances, and open follow-up items with evidence.

Finance, accounting, and procurement operationsPer variance
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
980 in · 1,183 out · 19.69s p50
$14.55 /100K
Current price · TrustedRouter
View token rates · observed 9/11/2026
TrustedRouter · $0.0211 input · $0.1055 output /1M
74%
Cheapest
DeepSeek V3
Hosted · Off-the-shelf
834 in · 1,156 out · 16.11s p50
$136 /100K
Best current price · OpenRouter
Compare 2 router rates · observed 9/11/2026
OpenRouter · $0.2500 input · $1.00 output /1M
TrustedRouter · $0.3376 input · $0.9390 output /1M
46%
GPT-5.4
Hosted · Off-the-shelf
799 in · 1,763 out · 16.83s p50
$2844 /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
88%
Best accuracy
All 12 model results and methodology

Generated 100/100 examples

ModelTierQualityJudgedScenario cost
Nemotron Nano 9B v2small4%100/100$51.09
Qwen3 235B A22Bmid16%100/100$361
DeepSeek V3mid46%100/100$136
Mistral Large 2407mid41%100/100$291
Arcee Trinity Large Thinkingmid16%100/100$162
GPT-5.4frontier88%100/100$2844
Claude Opus 4.7frontier32%100/100$5588
Gemini 3.1 Pro Previewfrontier1%100/100$2588
Gemma 4 E4B ITsmall74%100/100$14.55
Granite 4.1 8Bsmall13%100/100$18.31
Ministral 8B Instruct 2410small20%100/100$33.63
Qwen3 4B Instruct 2507small7%100/100$135

LLM-judge pass rate on 100 synthetic examples. Generator: gpt-5.2. Judge: gpt-5.2. Evaluated 2026-09-11T10:31:03.255Z.

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