Benchmark Almanac

FinGREAT

Claude Fable 5.1 92.8% leads since October 2026
Not in model cards

Models identify accounting relationships in financial passages. The authors compare 44 models against human annotations across 483 passages.

Released
October 2026
Kind
Finance
Level
University
Scored
Checked automatically
Answers
Short answer
Measures
Capability
Language
English
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
A host
In views
Law and finance
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
44
Entries in its paper, 35 models we track

2 models share the top score of 92.8% among models released by September 2026 2 models share the top score of 92.8% among models released by September 2026

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. This view shows only models we track, each at its best.

The newest model with a score here came out in September 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 Claude Fable 5.1 92.8%
  2. 2 GPT-6 Astra 92.8%
  3. 3 GPT-6.1 Sol 92.5%
  4. 4 Claude Opus 5.5 92.1%
  5. 5 Gemini 3.8 Flash 91.3%
  6. 6 Claude Sonnet 5.5 91.3%
  7. 7 Gemini 3.1 Pro 89.6%
  8. 8 Claude Haiku 4.5 84.5%
  9. 9 Qwen3.5-27B 84.1%
  10. 10 Qwen3.6 27B 84.1%
  11. 11 Gemma 4 12B 83.2%
  12. 12 Gemma 4 31B IT 83%
  13. 13 Gemini 3.5 Flash-Lite 82.4%
  14. 14 Ministral-3-14B-Instruct 80.1%
  15. 15 Gemma 4 26B A4B 79.9%
  16. 16 GPT-6 Luna 79.3%
  17. 17 Qwen3.5-9B 78.7%
  18. 18 Qwen3.5-35B-A3B 78.5%
  19. 19 Qwen 3.6 35B-A3B 77%
  20. 20 Gemma 3 27B 74.9%
  21. 21 Gemma 4 E4B 74.5%
  22. 22 Gemma 3 12B 71.2%
  23. 23 Phi-4 70.2%
  24. 24 Qwen3.5-4B 69.2%
  25. 25 Ministral 3 3B Instruct 66.9%
  26. 26 Gemma 4 E2B 65.2%
  27. 27 Ministral 3 8B Instruct 64.8%
  28. 28 Phi-3-Small-128K-Instruct 58.8%
  29. 29 Phi-3.5-mini 54.7%
  30. 30 Gemma 3 4B 54.2%
  31. 31 Llama-3.1-8B-Instruct 52.8%
  32. 32 Phi-3-mini-128k-instruct 49.5%
  33. 33 Llama-3.2-3B-Instruct 48%
  34. 34 Meta-Llama-3-8B-Instruct 46.6%
  35. 35 Phi-4 Mini 45.1%
  36. 36 DeepSeek-V2-Lite-Chat 42%
  37. 37 Qwen3.5-2B 41.8%
  38. 38 Qwen3.5-0.8B 36.4%
  39. 39 Llama-3.2-1B-Instruct 23%
  40. 40 Gemma 3 1B 22.4%
  41. 41 DeepSeek-LLM-7B-Chat 20.9%
  42. 42 Phi-3-medium-128k-instruct 19.5%
  43. 43 DeepSeek-MoE-16B-Chat 15.9%
  44. 44 Gemma 3 270M 8.90%

Source: The FinGREAT paper

Source: The FinGREAT paper

#EntryScoreSettingDateSource
1 Claude Fable 5.1 92.8% Sep 2026 Official leaderboard
2 GPT-6 Astra 92.8% Sep 2026 Official leaderboard
3 GPT-6.1 Sol 92.5% Sep 2026 Official leaderboard
4 Claude Opus 5.5 92.1% Sep 2026 Official leaderboard
5 Gemini 3.8 Flash 91.3% Sep 2026 Official leaderboard
6 Claude Sonnet 5.5 91.3% Sep 2026 Official leaderboard
7 Gemini 3.1 Pro 89.6% Feb 2026 Official leaderboard
8 Claude Haiku 4.5 84.5% Oct 2025 Official leaderboard
9 Qwen3.5-27B 84.1% Feb 2026 Official leaderboard
10 Qwen3.6 27B 84.1% Apr 2026 Official leaderboard
11 Gemma 4 12B 83.2% Jun 2026 Official leaderboard
12 Gemma 4 31B IT 83% Apr 2026 Official leaderboard
13 Gemini 3.5 Flash-Lite 82.4% Jul 2026 Official leaderboard
14 Ministral-3-14B-Instruct 80.1% Official leaderboard
15 Gemma 4 26B A4B 79.9% Apr 2026 Official leaderboard
16 GPT-6 Luna 79.3% Sep 2026 Official leaderboard
17 Qwen3.5-9B 78.7% Feb 2026 Official leaderboard
18 Qwen3.5-35B-A3B 78.5% Feb 2026 Official leaderboard
19 Qwen 3.6 35B-A3B 77% Apr 2026 Official leaderboard
20 Gemma 3 27B 74.9% Mar 2025 Official leaderboard
21 Gemma 4 E4B 74.5% Mar 2026 Official leaderboard
22 Gemma 3 12B 71.2% Mar 2025 Official leaderboard
23 Phi-4 70.2% Dec 2024 Official leaderboard
24 Qwen3.5-4B 69.2% Feb 2026 Official leaderboard
25 Ministral 3 3B Instruct 66.9% Dec 2025 Official leaderboard
26 Gemma 4 E2B 65.2% Mar 2026 Official leaderboard
27 Ministral 3 8B Instruct 64.8% Dec 2025 Official leaderboard
28 Phi-3-Small-128K-Instruct 58.8% Official leaderboard
29 Phi-3.5-mini 54.7% Aug 2024 Official leaderboard
30 Gemma 3 4B 54.2% Mar 2025 Official leaderboard
31 Llama-3.1-8B-Instruct 52.8% Official leaderboard
32 Phi-3-mini-128k-instruct 49.5% Apr 2024 Official leaderboard
33 Llama-3.2-3B-Instruct 48% Official leaderboard
34 Meta-Llama-3-8B-Instruct 46.6% Official leaderboard
35 Phi-4 Mini 45.1% Feb 2025 Official leaderboard
36 DeepSeek-V2-Lite-Chat 42% Official leaderboard
37 Qwen3.5-2B 41.8% Feb 2026 Official leaderboard
38 Qwen3.5-0.8B 36.4% Feb 2026 Official leaderboard
39 Llama-3.2-1B-Instruct 23% Official leaderboard
40 Gemma 3 1B 22.4% Mar 2025 Official leaderboard
41 DeepSeek-LLM-7B-Chat 20.9% Official leaderboard
42 Phi-3-medium-128k-instruct 19.5% May 2024 Official leaderboard
43 DeepSeek-MoE-16B-Chat 15.9% Official leaderboard
44 Gemma 3 270M 8.90% Aug 2025 Official leaderboard
#ModelScoreSettingDateSource
1 Claude Fable 5.1 92.8% Sep 2026 Official leaderboard
2 GPT-6 Astra 92.8% Sep 2026 Official leaderboard
3 GPT-6.1 Sol 92.5% Sep 2026 Official leaderboard
4 Claude Opus 5.5 92.1% Sep 2026 Official leaderboard
5 Gemini 3.8 Flash 91.3% Sep 2026 Official leaderboard
6 Claude Sonnet 5.5 91.3% Sep 2026 Official leaderboard
7 Gemini 3.1 Pro 89.6% Feb 2026 Official leaderboard
8 Claude Haiku 4.5 84.5% Oct 2025 Official leaderboard
9 Qwen3.5-27B 84.1% Feb 2026 Official leaderboard
10 Qwen3.6 27B 84.1% Apr 2026 Official leaderboard
11 Gemma 4 12B 83.2% Jun 2026 Official leaderboard
12 Gemma 4 31B IT 83% Apr 2026 Official leaderboard
13 Gemini 3.5 Flash-Lite 82.4% Jul 2026 Official leaderboard
14 Gemma 4 26B A4B 79.9% Apr 2026 Official leaderboard
15 GPT-6 Luna 79.3% Sep 2026 Official leaderboard
16 Qwen3.5-9B 78.7% Feb 2026 Official leaderboard
17 Qwen3.5-35B-A3B 78.5% Feb 2026 Official leaderboard
18 Qwen 3.6 35B-A3B 77% Apr 2026 Official leaderboard
19 Gemma 3 27B 74.9% Mar 2025 Official leaderboard
20 Gemma 4 E4B 74.5% Mar 2026 Official leaderboard
21 Gemma 3 12B 71.2% Mar 2025 Official leaderboard
22 Phi-4 70.2% Dec 2024 Official leaderboard
23 Qwen3.5-4B 69.2% Feb 2026 Official leaderboard
24 Ministral 3 3B Instruct 66.9% Dec 2025 Official leaderboard
25 Gemma 4 E2B 65.2% Mar 2026 Official leaderboard
26 Ministral 3 8B Instruct 64.8% Dec 2025 Official leaderboard
27 Phi-3.5-mini 54.7% Aug 2024 Official leaderboard
28 Gemma 3 4B 54.2% Mar 2025 Official leaderboard
29 Phi-3-mini-128k-instruct 49.5% Apr 2024 Official leaderboard
30 Phi-4 Mini 45.1% Feb 2025 Official leaderboard
31 Qwen3.5-2B 41.8% Feb 2026 Official leaderboard
32 Qwen3.5-0.8B 36.4% Feb 2026 Official leaderboard
33 Gemma 3 1B 22.4% Mar 2025 Official leaderboard
34 Phi-3-medium-128k-instruct 19.5% May 2024 Official leaderboard
35 Gemma 3 270M 8.90% Aug 2025 Official leaderboard

Rankings by part

We read 14 parts from the benchmark’s paper, each ranking models on one part of the test, such as Finance, Investment and Operations. Claude Fable 5.1 leads 5 of them, and 9 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Finance

35 models we track have a score here.

  1. 1 Qwen3.5-2B ↑30 89%
  2. 2 Qwen3.5-0.8B ↑30 86.6%
  3. 3 Claude Fable 5.1 ↓2 84.1%
  4. 4 GPT-6 Astra ↓2 84.1%
  5. 5 Claude Opus 5.5 ↓1 84.1%
  6. 6 GPT-6.1 Sol ↓3 84.1%
  7. 7 Claude Sonnet 5.5 ↓1 80.5%
  8. 8 Gemini 3.8 Flash ↓3 79.3%
  9. 9 Gemini 3.1 Pro ↓2 78%
  10. 10 Claude Haiku 4.5 ↓2 72%

Number one over time

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

Anthropic
Benchmark paper Passed 85% in Oct 2026, by Claude Fable 5.1

Source: The FinGREAT paper

Gemini 3.1 Pro on the official leaderboard passed 85% in February 2026

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
In the benchmark’s paper Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Claude Fable 5.1 Gemini 3.1 Pro Claude Haiku 4.5 Gemma 3 27B Phi-4 Phi-3.5-mini Phi-3-mini-128k-instruct 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Claude Fable 5.1 Claude Haiku 4.5

Source: Official FinGREAT 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 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 4 pass 1 note
  • Pass Independent runs Scores from the benchmark’s paper (44 entries).
  • Pass Models scored 35 models we track, plus 9 other entries on its leaderboard.
  • Pass Newest score The newest is from September 2026, for GPT-6.1 Sol.
  • Pass Frontier models 5 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Note Headroom Saturated. The best is 92.8%, past 85% since February 2026.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 35 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