---
type: "post"
id: "artificial-analysis-launched-the-decision-index-v1-0-5ce031"
slug: "artificial-analysis-launched-the-decision-index-v1-0-5ce031"
title: "Artificial Analysis launched the Decision Index v1.0 , benchmarking decision models on classification, tool calling, knowledge, and reasoning."
description: "Thesis: Decision-model evaluation becomes more useful when it treats bounded choices as a separate task class and reports more than aggregate accuracy."
retrieval_nugget: "Thesis: Decision-model evaluation becomes more useful when it treats bounded choices as a separate task class and reports more than aggregate accuracy."
published_at: "2026-10-14"
updated_at: "2026-10-14"
record_date: "2026-10-14"
date_kind: "published_at"
topics: []
source_urls: ["https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:artificial-analysis-launched-the-decision-index-v1-0-5ce031"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:6ef9c17cf91f2c51ccad13e412cc9bd737174f8e59466fbcb4845500fe236a8c","reviewed_at":"2026-10-14T02:01:43.422Z","source_evidence_count":1,"verified_claim_count":2,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":2,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Decision-model evaluation becomes more useful when it treats bounded choices as a separate task class and reports more than aggregate accuracy.","observed_facts":[{"text":"The published index combines AA-Omniscience-Decision, HLE-MultiChoice, When2Call, and WildGuardTest, assigning each a 25% weight.","source_urls":["https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index"]},{"text":"The evaluation page reports calibration, median time per task, and estimated cost per million tasks.","source_urls":["https://artificialanalysis.ai/evaluations/artificial-analysis-decision-index"]}],"mechanism":"Decision models select among fixed options and provide a probability for each option, making calibration and operational cost relevant alongside correctness.","why_now":"Builders increasingly use models for routing, classification, and tool selection, where outputs are constrained choices rather than open-ended prose.","implications":["Teams can use the index as a screening comparison, then test candidate models against their own task distribution and acceptance criteria.","Calibration, latency, and cost are relevant comparison dimensions in addition to aggregate correctness."],"evidence_boundary":"The source describes the benchmark and its displayed metrics; it does not establish external validation, representativeness for a given deployment, or causal links between index rank and business outcomes.","watch_conditions":["Whether independent evaluations or production reports show that index scores predict performance on real bounded-decision workloads."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied; this creates a distinct analytical record for the newly launched decision-model benchmark."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/artificial-analysis-launched-the-decision-index-v1-0-5ce031/","markdown":"https://newruntime.com/posts/artificial-analysis-launched-the-decision-index-v1-0-5ce031.md","json":"https://newruntime.com/posts/artificial-analysis-launched-the-decision-index-v1-0-5ce031.json"}
---

# Artificial Analysis launched the Decision Index v1.0 , benchmarking decision models on classification, tool calling, knowledge, and reasoning.

## Retrieval answer

Thesis: Decision-model evaluation becomes more useful when it treats bounded choices as a separate task class and reports more than aggregate accuracy.

Thesis: Decision-model evaluation becomes more useful when it treats bounded choices as a separate task class and reports more than aggregate accuracy. The Artificial Analysis Decision Index is a new composite evaluation, not evidence that a benchmark score alone predicts production success.

Observed facts: the published index combines four evaluations—AA-Omniscience-Decision, HLE-MultiChoice, When2Call, and WildGuardTest—and assigns each a 25% weight. The page also reports calibration, median time per task, and estimated cost per million tasks. Mechanism: Decision models select among fixed options and provide a probability for each option, making calibration and operational cost relevant alongside correctness.

Why now: Builders increasingly use models for routing, classification, and tool selection, where outputs are constrained choices rather than open-ended prose. First implication: Teams can use the index as a screening comparison, then test candidate models against their own task distribution and acceptance criteria. Evidence boundary: The source describes the benchmark and its displayed metrics; it does not establish external validation, representativeness for a given deployment, or causal links between index rank and business outcomes. First watch condition: Whether independent evaluations or production reports show that index scores predict performance on real bounded-decision workloads.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)