docketrouter
Models / Qwen

Qwen3 30B A3B Instruct 2507

by Qwen · qwen/qwen3-30b-a3b-instruct-2507

Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter mixture-of-experts language model from Qwen, with 3.3B active parameters per inference. It operates in non-thinking mode and is designed for high-quality instruction following, multilingual understanding, and...

tool-usereleased 2025-07-29
Legal score · raw
-
not yet benchmarked
Context
262K
max output 32K
Input
$0.05
per 1M tokens
Output
$0.19
per 1M tokens
Suite cost
-
run the suite to see

Benchmark results

TaskCategoryRawJuicedCorrectLatencyCostRan
Hearsay IdentificationEvidence------
Bluebook Citation FormatResearch & Writing------
Federal Civil ProcedureProcedure------
Limitations ArithmeticProcedure------
Contract Clause ClassificationContracts------
Citation Hallucination ResistanceReliability------

Measured by DocketBuster

These numbers come from DocketBuster's own legal battery, not from DocketRouter's suite. Latest run per metric, with n and a 95% Wilson interval where the source reports one. See docketbuster.com/benchmarks.

MetricValuenIntervalMeasured
Statute pinpoint, exact section (no retrieval)7.0% (21/299)29995% CI 4.6% to 10.5%2026-08-21
Statute pinpoint, exact section (with DocketBuster retrieval)76.9% (236/307)30795% CI 71.8% to 81.2%2026-08-24
Say-nothing rate (declines to bluff when the answer is not in the record)99.7%28695% CI 98.0% to 99.9%2026-08-21
Abstained on statute pinpoint4.6% (14/307)307count, no interval reported2026-08-24
Coaching quality (GW-14x, 0 to 8)7.36 / 8-rubric mean, no interval reported2026-08-22

Source files: hard-llm-ortier-cheap-statute.json, hard-llm-ortier-cheap-say.json, gw14x-ortier-cheap.json, hard-llm-level-qwen3-30b-a3b-statute_rag.json. Raw model name in source: qwen/qwen3-30b-a3b-instruct-2507.