Benchmark Almanac

Financial Coaching Bench

Claude Opus 5.5 76.2% leads since November 2026
Not in model cards

Models interview a simulated client about a household financial problem and propose a plan.

The judge checks the plan against practitioner criteria across 256 scored scenarios.

Made by
Vals AI
Released
November 2026
Kind
Finance
Level
Expert
Scored
Graded by a model
Answers
Agent task
Tasks
Kept private
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Multi-turn
Items from
Generated
Refresh
One fixed set
Run by
A host
Occupations
Business and financial operations
In views
Law and finance
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
October 2026
Scores refreshed
November 20th, 2026
27
Models scored on its own leaderboard

Claude Opus 5.5 leads with 76.2%

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.

  1. 1 Claude Opus 5.5 76.2%
  2. 2 Muse Spark 1.2 75.7%
  3. 3 Claude Fable 5.1 75.7%
  4. 4 Qwen 3.8 Max 75.6%
  5. 5 MiMo-V2.6 Pro 75.4%
  6. 6 Muse Wonder Full 75%
  7. 7 Claude Sonnet 5.5 74.9%
  8. 8 GPT-5.6 Sol 74.9%
  9. 9 Kimi K3 74.5%
  10. 10 Grok 4.7 74.1%
  11. 11 GLM-5.3 74.1%
  12. 12 Muse Wonder 73.8%
  13. 13 Gemini 4 Argon 73.1%
  14. 14 DeepSeek V4.1 Flash 72.8%
  15. 15 Muse Spark 1.3 71%
  16. 16 GPT-5.6 Terra 71%
  17. 17 GPT-5.6 Luna 69.3%
  18. 18 MiniMax-M3 69%
  19. 19 GPT-6.1 Sol 68.5%
  20. 20 GPT-6 Astra 67.5%
  21. 21 Gemini 3.1 Pro 66.7%
  22. 22 Gemini 3.8 Flash 66.3%
  23. 23 Gemini 3.6 Flash 65.8%
  24. 24 GPT-6 Luna 65.2%
  25. 25 Gemini 3.5 Flash 63%
  26. 26 GPT-6 Sol 62.6%
  27. 27 Gemini 3.5 Flash-Lite 57.5%

Source: Official Financial Coaching Bench leaderboard

#ModelScore ±SettingDateSource
1 Claude Opus 5.5 76.2% ±3 Sep 2026 Official leaderboard
2 Muse Spark 1.2 75.7% ±3 Aug 2026 Official leaderboard
3 Claude Fable 5.1 75.7% ±3 Sep 2026 Official leaderboard
4 Qwen 3.8 Max 75.6% ±2.9 Aug 2026 Official leaderboard
5 MiMo-V2.6 Pro 75.4% ±2.8 Sep 2026 Official leaderboard
6 Muse Wonder Full 75% ±3.1 Oct 2026 Official leaderboard
7 Claude Sonnet 5.5 74.9% ±3 Sep 2026 Official leaderboard
8 GPT-5.6 Sol 74.9% ±3 Jul 2026 Official leaderboard
9 Kimi K3 74.5% ±2.9 Jul 2026 Official leaderboard
10 Grok 4.7 74.1% ±3 Sep 2026 Official leaderboard
11 GLM-5.3 74.1% ±2.9 Aug 2026 Official leaderboard
12 Muse Wonder 73.8% ±3 Oct 2026 Official leaderboard
13 Gemini 4 Argon 73.1% ±2.9 Sep 2026 Official leaderboard
14 DeepSeek V4.1 Flash 72.8% ±3.1 Sep 2026 Official leaderboard
15 Muse Spark 1.3 71% ±3 Sep 2026 Official leaderboard
16 GPT-5.6 Terra 71% ±3 Jul 2026 Official leaderboard
17 GPT-5.6 Luna 69.3% ±3.2 Jul 2026 Official leaderboard
18 MiniMax-M3 69% ±3.1 Jun 2026 Official leaderboard
19 GPT-6.1 Sol 68.5% ±3.1 Sep 2026 Official leaderboard
20 GPT-6 Astra 67.5% ±3.3 Sep 2026 Official leaderboard
21 Gemini 3.1 Pro 66.7% ±2.9 Feb 2026 Official leaderboard
22 Gemini 3.8 Flash 66.3% ±3.1 Sep 2026 Official leaderboard
23 Gemini 3.6 Flash 65.8% ±3.1 Jul 2026 Official leaderboard
24 GPT-6 Luna 65.2% ±3.3 Sep 2026 Official leaderboard
25 Gemini 3.5 Flash 63% ±3.2 May 2026 Official leaderboard
26 GPT-6 Sol 62.6% ±3.4 Sep 2026 Official leaderboard
27 Gemini 3.5 Flash-Lite 57.5% ±3.3 Jul 2026 Official leaderboard

The best score, 76.2%, costs $1.63 a task

Runs on the cost frontier score higher than any run that's cheaper. MiMo-V2.6 Pro comes within 0.8 points of the best for $0.01 a task, 163 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.

Meta Xiaomi Corp Anthropic Cost frontier One model at each effort level
0% 20% 40% 60% 80% 100% $0.01 $0.10 $1.00 $10 Cost per task, log scale Claude Opus 5.5 Muse Spark 1.2 MiMo-V2.6 Pro 0% 20% 40% 60% 80% 100% $0.01 $0.10 $1.00 $10 Cost per task, log scale Claude Opus 5.5 MiMo-V2.6 Pro

Source: Official Financial Coaching Bench leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Claude Opus 5.5 76.2% $1.63
Muse Spark 1.2 75.7% $0.06
Claude Fable 5.1 75.7% $1.14
Qwen 3.8 Max 75.6% $0.13
MiMo-V2.6 Pro 75.4% $0.01
Muse Wonder Full 75% $0.44
Claude Sonnet 5.5 74.9% $0.80
GPT-5.6 Sol 74.9% $0.47
Kimi K3 74.5% $0.19
Grok 4.7 74.1% $0.24
GLM-5.3 74.1% $0.09
Muse Wonder 73.8% $0.32
Gemini 4 Argon 73.1% $0.16
DeepSeek V4.1 Flash 72.8% $0.02
Muse Spark 1.3 71% $0.07
GPT-5.6 Terra 71% $0.31
Muse Spark 1.3 max 70.3% $0.09
GPT-5.6 Luna 69.3% $0.02
MiniMax-M3 69% $0.03
GPT-6.1 Sol 68.5% $0.30
GPT-6 Astra 67.5% $1.72
Gemini 3.1 Pro 66.7% $0.12
Gemini 3.8 Flash 66.3% $0.10
Gemini 3.6 Flash 65.8% $0.10
GPT-6 Luna 65.2% $0.02
Gemini 3.5 Flash 63% $0.13
GPT-6 Sol 62.6% $0.23
Gemini 3.5 Flash-Lite 57.5% $0.03

Number one over time

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

Anthropic
Own leaderboard 1 leader

Source: Official Financial Coaching Bench 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%) Claude Opus 5.5 GPT-5.6 Sol MiniMax-M3 Gemini 3.1 Pro 0% 20% 40% 60% 80% 100% 2026 Near the ceiling (85%) Claude Opus 5.5 MiniMax-M3

Source: Official Financial Coaching Bench 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

All 11 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access The tasks are private.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (27 entries).
  • Pass Models scored 27 models we track.
  • Pass Newest score The newest is from October 2026, for Muse Wonder.
  • Pass Frontier models 6 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Pass Headroom The best is 76.2%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on vals.ai: 28 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