Benchmark Almanac

AMLBench v0.2 (Superjective)

Llama 3-70B 100% leads since July 2026
Not in model cards

Models decide whether to escalate transaction alerts. The score measures how often they escalate reportable alerts under changes to analyst instructions and case narratives.

The authors tested seven models. Six qualify for the leaderboard by clearing at least 11 of 12 benign alerts. The sample also includes 48 reportable alerts.

Released
July 2026
Kind
Finance
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Run by
Anyone
Occupations
Business and financial operations
Strongest model tested
GPT-5.5 #21 on the capability index
Last model added
May 2026
Scores refreshed
November 20th, 2026
6
Models scored on its own leaderboard

5 models share the top score of 100% among models released by May 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 May 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 Llama 3-70B Baseline (no stressor) 100%
  2. 2 GPT-4o Mar 2025, Baseline (no stressor) 100%
  3. 3 GPT-5.5 Baseline (no stressor) 100%
  4. 4 Grok 4.3 Baseline (no stressor) 100%
  5. 5 Claude Opus 4.8 Baseline (no stressor) 100%
  6. 6 Claude Haiku 4.5 Baseline (no stressor) 97.9%

Source: Official AMLBench v0.2 (Superjective) leaderboard

#ModelScoreSettingDateSource
1 Llama 3-70B 100% Baseline (no stressor) Apr 2024 Official leaderboard
2 GPT-4o Mar 2025 100% Baseline (no stressor) Mar 2025 Official leaderboard
3 GPT-5.5 100% Baseline (no stressor) Apr 2026 Official leaderboard
4 Grok 4.3 100% Baseline (no stressor) Apr 2026 Official leaderboard
5 Claude Opus 4.8 100% Baseline (no stressor) May 2026 Official leaderboard
6 Claude Haiku 4.5 97.9% Baseline (no stressor) Oct 2025 Official leaderboard

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
Llama 3-70B
+ deceptive… + prompt at… Baseline (n…
100%
GPT-4o
+ deceptive… + prompt at… Baseline (n…
100%
GPT-5.5
+ deceptive… Baseline (n…
100%
Grok 4.3
+ deceptive… + prompt at… Baseline (n…
100%
Claude Opus 4.8
+ deceptive… + prompt at…
100%
Claude Haiku 4.5
+ deceptive… + prompt at…
97.9%

Source: Official AMLBench v0.2 (Superjective) leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Llama 3-70B Baseline (no stressor) 100%
GPT-4o Mar 2025 Baseline (no stressor) 100%
GPT-5.5 Baseline (no stressor) 100%
GPT-5.5 + prompt attack 100%
Grok 4.3 Baseline (no stressor) 100%
Claude Opus 4.8 Baseline (no stressor) 100%
Claude Haiku 4.5 Baseline (no stressor) 97.9%
Claude Opus 4.8 + prompt attack 95.8%
GPT-4o Mar 2025 + prompt attack 93.8%
Claude Haiku 4.5 + prompt attack 93.8%
Llama 3-70B + prompt attack 87.5%
Claude Opus 4.8 + deceptive cover (A2) 72.9%
GPT-5.5 + deceptive cover (A2) 54.2%
Grok 4.3 + prompt attack 52.1%
Grok 4.3 + deceptive cover (A2) 43.8%
Claude Haiku 4.5 + deceptive cover (A2) 32.9%
Llama 3-70B + deceptive cover (A2) 31.2%
GPT-4o Mar 2025 + deceptive cover (A2) 27.1%

Number one over time

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

Meta
Own leaderboard Passed 85% in Jul 2026, by Llama 3-70B

Source: Official AMLBench v0.2 (Superjective) leaderboard

Llama 3-70B on the official leaderboard passed 85% in April 2024

Models released before it came out already scored that high. 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
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Llama 3-70B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Llama 3-70B

Source: Official AMLBench v0.2 (Superjective) 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (6 entries).
  • Pass Models scored 6 models we track.
  • Note Newest score The newest came 6 months ago, in May 2026, for Claude Opus 4.8. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5.5 (#21).
  • Note Headroom Saturated. The best is 100%, past 85% since April 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on amlbench.com: 18 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