Benchmark Almanac

FinDeepIndicator

Qwen 3.6 Max (Preview) 39% leads since August 2026
Not in model cards

Agents construct financial indicators by specifying formulas, retrieving market data, and calculating results. The evaluation checks performance at each step.

Made by
Chaoqun Yang and colleagues
Released
August 2026
Kind
Finance
Level
University
Scored
Checked automatically
Answers
Agent task
Measures
Capability
Language
English
In and out
Text, Time series, Tables → Actions, Text
Interaction
Agentic
Environment
Browser
Items from
Real work and use
Refresh
One fixed set
Run by
A host
Strongest model tested
Claude Sonnet 4.6 #57 on the capability index
Last model added
April 2026
Scores refreshed
November 20th, 2026
7
Models scored on its own leaderboard

Qwen 3.6 Max (Preview) leads with 39% among models released by April 2026

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart.

The newest model with a score here came out in April 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 Qwen 3.6 Max (Preview) Deep Research 39%
  2. 2 Claude Sonnet 4.6 Deep Research 36.6%
  3. 3 Gemini 3 Flash Deep Research 20.6%
  4. 4 DeepSeek-V4-Flash LLM with Search 8.00%
  5. 5 GPT-5 LLM with Search 6.30%
  6. 6 Grok 4.3 LLM with Search 1.70%
  7. 7 GPT-5 mini LLM with Search 1.50%

Source: Official FinDeepIndicator leaderboard

#ModelScoreSettingSubmitted byReleasedRun onSource
1 Qwen 3.6 Max (Preview) 39% Deep Research Official Apr 2026 Aug 2026 Official leaderboard
2 Claude Sonnet 4.6 36.6% Deep Research Official Feb 2026 Aug 2026 Official leaderboard
3 Gemini 3 Flash 20.6% Deep Research Official Dec 2025 Aug 2026 Official leaderboard
4 DeepSeek-V4-Flash 8.00% LLM with Search Official Apr 2026 Aug 2026 Official leaderboard
5 GPT-5 6.30% LLM with Search Official Aug 2025 Aug 2026 Official leaderboard
6 Grok 4.3 1.70% LLM with Search Official Apr 2026 Aug 2026 Official leaderboard
7 GPT-5 mini 1.50% LLM with Search Official Aug 2025 Aug 2026 Official leaderboard

Rankings by part

We read 2 parts from the benchmark’s own leaderboard: U.S. market and Chinese market. Each ranks models on one part of the test. Qwen 3.6 Max (Preview) leads both. Arrows show how far a model moved from its overall rank among the models we track.

U.S. market

7 models we track have a score here.

  1. 1 Qwen 3.6 Max (Preview) 40.5%
  2. 2 Claude Sonnet 4.6 36.5%
  3. 3 Gemini 3 Flash 25.4%
  4. 4 GPT-5 ↑1 7.90%
  5. 5 DeepSeek-V4-Flash ↓1 7.20%
  6. 6 Grok 4.3 2.30%
  7. 7 GPT-5 mini 1.80%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Qwen 3.6 Max (Preview)
LLM with Se… Deep Resear…
39%
Claude Sonnet 4.6
LLM with Se… Deep Resear…
36.6%
Gemini 3 Flash
LLM with Se… Deep Resear…
20.6%

Source: Official FinDeepIndicator leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Qwen 3.6 Max (Preview) Deep Research 39%
Claude Sonnet 4.6 Deep Research 36.6%
Gemini 3 Flash Deep Research 20.6%
Gemini 3 Flash LLM with Search 11.8%
Claude Sonnet 4.6 LLM with Search 8.60%
Qwen 3.6 Max (Preview) LLM with Search 8.10%
DeepSeek-V4-Flash LLM with Search 8.00%
GPT-5 LLM with Search 6.30%
Grok 4.3 LLM with Search 1.70%
GPT-5 mini LLM with Search 1.50%

Number one over time

Lab claims count from the day the card came out, independent runs from the model’s release.

Qwen
Own leaderboard Not tested since Aug 2026

Source: Official FinDeepIndicator leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Qwen 3.6 Max (Preview) Claude Sonnet 4.6 Gemini 3 Flash GPT-5 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Qwen 3.6 Max (Preview) Gemini 3 Flash

Source: Official FinDeepIndicator leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Law, finance and medicine

All 637 in Law, finance and medicine
  • MedXpertQA-MM In use Models answer expert-level medical exam questions that combine images with clinical records and test results. Scoring is accuracy. Qwen 3.8 Max 80.4%
  • HealthBench Professional In use 525 physician-authored tasks from real clinician chats (care consult, documentation, research) with physician-adjudicated rubrics. Claude Sonnet 5.5 69.2%
  • HealthBench In use 5,000 realistic health conversations graded against physician-written rubrics. A length-adjusted variant penalizes verbosity. GPT-5 67.2%
  • FinSearchComp In use Agents search the web to fetch live financial data, look up history, and run multi-step investigations, across 635 expert-written questions. Scored by answer accuracy. Ling-3.1-flash 81.7%
  • KorMedMCQA In use Models answer 7,469 multiple-choice questions from Korean licensing exams for doctors, nurses, pharmacists and dentists, held from 2012 to 2024. o1-preview 92.7%
  • Harvey LAB-AA v1.1 In use Artificial Analysis's run of Harvey's Legal Agent Benchmark on Harvey's newer private set of 120 tasks across 24 practice areas. A deliverable with a material hallucination scores zero, and the headline is the share of judges who find every rubric criterion passed. GPT-6.2 Astra 15%

Data audit

8 of 10 checks pass, 2 warnings. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (7 entries).
  • Pass Models scored 7 models we track.
  • Warning Newest score The newest came 7 months ago, in April 2026, for Grok 4.3.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude Sonnet 4.6 (#57).
  • Pass Headroom The best is 39%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on openfinarena.com: 10 rows for models we track at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check