Benchmark Almanac

Fin-Fact

GPT4 (snapshot unspecified) 76% tops its leaderboard Claude 3 Opus 61% leads the models we track
Not in model cards

Models judge whether evidence supports a financial claim, contradicts it, or leaves too little information. The original evaluation also tests models with images.

The original authors tested GPT-4V and Gemini Vision Pro. Suprabhat Rapolu independently reports macro F1 on 3,367 scored pairs.

Released
September 2023
Kind
Finance
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
English
In and out
Text, Images → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Claude 3 Opus #173 on the capability index
Last model added
February 2024
Scores refreshed
November 20th, 2026
12
Entries in its paper, 3 models we track

GPT4 (snapshot unspecified) tops the leaderboard with 76% Claude 3 Opus leads with 61% among models released by February 2024

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Claude 3 Opus, at 61%. This view shows only models we track, each at its best.

The newest model with a score here came out in February 2024, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT4 (snapshot unspecified) Text 76%
  2. 2 Claude 3 Opus Text 61%
  3. 3 Gemini Vision Pro Text, Image 59%
  4. 4 GPT4-Vision Text, Image 49%
  5. 5 Mixtral 8x7B Text 48%
  6. 6 LLaVA(13b) Text, Image 45%
  7. 7 Gemini 1.0 Pro Text 44%
  8. 8 BART-Large Text 34.4%
  9. 9 GPT-2 Text 31.2%
  10. 10 InstructBLIP Text, Image 30%
  11. 11 ELECTRA Text 28.6%
  12. 12 RoBERTa-Large Text 25.5%

Source: The Fin-Fact paper

#EntryScoreSettingDateSource
1 GPT4 (snapshot unspecified) 76% Text Official leaderboard
2 Claude 3 Opus 61% Text Feb 2024 Official leaderboard
3 Gemini Vision Pro 59% Text, Image Official leaderboard
4 GPT4-Vision 49% Text, Image Official leaderboard
5 Mixtral 8x7B 48% Text Dec 2023 Official leaderboard
6 LLaVA(13b) 45% Text, Image Official leaderboard
7 Gemini 1.0 Pro 44% Text Dec 2023 Official leaderboard
8 BART-Large 34.4% Text Official leaderboard
9 GPT-2 31.2% Text Official leaderboard
10 InstructBLIP 30% Text, Image Official leaderboard
11 ELECTRA 28.6% Text Official leaderboard
12 RoBERTa-Large 25.5% Text Official leaderboard
#ModelScoreSettingDateSource
1 Claude 3 Opus 61% Text Feb 2024 Official leaderboard
2 Mixtral 8x7B 48% Text Dec 2023 Official leaderboard
3 Gemini 1.0 Pro 44% Text Dec 2023 Official leaderboard

Rankings by part

We read 3 parts from the benchmark’s paper: Precision, Recall and Accuracy. Each ranks models on one part of the test. Claude 3 Opus leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.

Precision

3 models we track have a score here.

  1. 1 Claude 3 Opus 62%
  2. 2 Gemini 1.0 Pro ↑1 45%
  3. 3 Mixtral 8x7B ↓1 42%

2 models have held number one in the benchmark’s paper since its release in September 2023

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

Mistral AI Anthropic
Benchmark paper Not tested since Feb 2024

Source: The Fin-Fact paper

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
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) Claude 3 Opus Mixtral 8x7B 0% 20% 40% 60% 80% 100% 2024 Near the ceiling (85%) Claude 3 Opus Mixtral 8x7B

Source: Official Fin-Fact 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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 2 years ago, in February 2024, for Claude 3 Opus. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (12 entries).
  • Pass Models scored 3 models we track, plus 9 other entries on its leaderboard.
  • Warning Newest score The newest came 2 years ago, in February 2024, for Claude 3 Opus.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Claude 3 Opus (#173).
  • Pass Headroom The best is 61%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 3 rows for models we track and 9 other entries 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