AgenticFinBench
Agents complete 230 financial requests in two languages. Graders apply task-specific rubrics and audit unsupported claims against external evidence.
- Made by
- NashNova Research
- Released
- November 2026
- Kind
- Finance
- Level
- Expert
- Scored
- Graded by a model
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Subject
- Economics and business
- Interaction
- Agentic
- Environment
- APIs and tools
- Refresh
- One fixed set
- Run by
- Anyone
- In views
- Agents, Law and finance
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- September 2026
- Scores refreshed
- November 20th, 2026
GPT-5.6 Sol leads with 89.2 among models released by September 2026 GPT-5.6 Sol leads with 89.2 among models released by September 2026
Best score per entry 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. 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:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6.1 Terra
- and 11 more
- 1 GPT-5.6 Sol 89.2
- 2
DeepSeek V4.1 Flash 88.8
- 3 GPT-6.1 Sol 87.7
- 4 GPT-6 Astra 87.3
- 5 Kimi K3 86.7
- 6
Gemini 4 Argon 85.1
- 7 Claude Opus 5.5 84.9
- 8 GPT-6 Luna 78.7
- 9
Seed2.1 Pro 78.0
- 10 DuMate 65.4
Source: Official AgenticFinBench leaderboard
- 1 GPT-5.6 Sol 89.2
- 2
DeepSeek V4.1 Flash 88.8
- 3 GPT-6.1 Sol 87.7
- 4 GPT-6 Astra 87.3
- 5 Kimi K3 86.7
- 6
Gemini 4 Argon 85.1
- 7 Claude Opus 5.5 84.9
- 8 GPT-6 Luna 78.7
- 9
Seed2.1 Pro 78.0
Source: Official AgenticFinBench leaderboard
| # | Entry | Score ± | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.6 Sol | 89.2 87.6 to 90.7 | Codex CLI | Jul 2026 | Official leaderboard |
| 2 | | 88.8 ±1.3 | Codex CLI | Sep 2026 | Official leaderboard |
| 3 | GPT-6.1 Sol | 87.7 86 to 89.3 | Codex CLI | Sep 2026 | Official leaderboard |
| 4 | GPT-6 Astra | 87.3 ±1.6 | Codex CLI | Sep 2026 | Official leaderboard |
| 5 | Kimi K3 | 86.7 84.4 to 88.9 | Kimi hosted | Jul 2026 | Official leaderboard |
| 6 | | 85.1 83.3 to 86.8 | Gemini CLI | Sep 2026 | Official leaderboard |
| 7 | Claude Opus 5.5 | 84.9 ±1.5 | Pi | Sep 2026 | Official leaderboard |
| 8 | GPT-6 Luna | 78.7 ±2 | Codex CLI | Sep 2026 | Official leaderboard |
| 9 | | 78.0 74.7 to 81 | Trae Work | Jun 2026 | Official leaderboard |
| 10 | DuMate | 65.4 ±2.9 | DuMate CLI | Official leaderboard |
| # | Model | Score ± | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5.6 Sol | 89.2 87.6 to 90.7 | Codex CLI | Jul 2026 | Official leaderboard |
| 2 | | 88.8 ±1.3 | Codex CLI | Sep 2026 | Official leaderboard |
| 3 | GPT-6.1 Sol | 87.7 86 to 89.3 | Codex CLI | Sep 2026 | Official leaderboard |
| 4 | GPT-6 Astra | 87.3 ±1.6 | Codex CLI | Sep 2026 | Official leaderboard |
| 5 | Kimi K3 | 86.7 84.4 to 88.9 | Kimi hosted | Jul 2026 | Official leaderboard |
| 6 | | 85.1 83.3 to 86.8 | Gemini CLI | Sep 2026 | Official leaderboard |
| 7 | Claude Opus 5.5 | 84.9 ±1.5 | Pi | Sep 2026 | Official leaderboard |
| 8 | GPT-6 Luna | 78.7 ±2 | Codex CLI | Sep 2026 | Official leaderboard |
| 9 | | 78.0 74.7 to 81 | Trae Work | Jun 2026 | Official leaderboard |
The best score, 89.2, costs $0.31 a task
Runs on the cost frontier score higher than any run that's cheaper. DeepSeek V4.1 Flash comes within 0.4 of the best for $0.02 a task, 16 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
Source: Official AgenticFinBench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-5.6 Sol | Codex CLI | 89.2 | $0.31 | ||
| Codex CLI | 88.8 | $0.02 | |||
| GPT-5.6 Sol | Pi | 88.6 | $0.72 | ||
| Pi | 87.8 | $0.06 | |||
| GPT-6.1 Sol | Codex CLI | 87.7 | $0.17 | ||
| GPT-6 Astra | Codex CLI | 87.3 | $0.56 | ||
| Kimi K3 | Kimi hosted | 86.7 | $0.35 | ||
| DSH | 85.8 | $0.07 | |||
| Gemini CLI | 85.1 | $0.22 | |||
| Claude Opus 5.5 | Pi | 84.9 | $0.20 | ||
| Claude Opus 5.5 | Claude Code CLI | 84.4 | $0.17 | ||
| GPT-5.6 Sol | DSH | 84.0 | $0.91 | ||
| Claude Opus 5.5 | Codex CLI | 83.1 | $0.28 | ||
| Claude Opus 5.5 | DSH | 80.7 | $0.15 | ||
| GPT-6 Luna | Codex CLI | 78.7 | $0.008 | ||
| Trae Work | 78.0 | $0.31 | |||
| WorkBuddy | 71.3 | $0.03 |
The best score, 89.2, costs $0.31 a task
Runs on the cost frontier score higher than any run that's cheaper. DeepSeek V4.1 Flash comes within 0.4 of the best for $0.02 a task, 16 times cheaper. The costs are the leaderboard's own estimates, so compare them only within this benchmark.
Source: Official AgenticFinBench leaderboard
| Model | Effort | Harness | Score | Cost per task | Output tokens |
|---|---|---|---|---|---|
| GPT-5.6 Sol | Codex CLI | 89.2 | $0.31 | ||
| Codex CLI | 88.8 | $0.02 | |||
| GPT-5.6 Sol | Pi | 88.6 | $0.72 | ||
| Pi | 87.8 | $0.06 | |||
| GPT-6.1 Sol | Codex CLI | 87.7 | $0.17 | ||
| GPT-6 Astra | Codex CLI | 87.3 | $0.56 | ||
| Kimi K3 | Kimi hosted | 86.7 | $0.35 | ||
| DSH | 85.8 | $0.07 | |||
| Gemini CLI | 85.1 | $0.22 | |||
| Claude Opus 5.5 | Pi | 84.9 | $0.20 | ||
| Claude Opus 5.5 | Claude Code CLI | 84.4 | $0.17 | ||
| GPT-5.6 Sol | DSH | 84.0 | $0.91 | ||
| Claude Opus 5.5 | Codex CLI | 83.1 | $0.28 | ||
| Claude Opus 5.5 | DSH | 80.7 | $0.15 | ||
| GPT-6 Luna | Codex CLI | 78.7 | $0.008 | ||
| Trae Work | 78.0 | $0.31 | |||
| WorkBuddy | 71.3 | $0.03 |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official AgenticFinBench leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official AgenticFinBench 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
- 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 10 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
- Pass Independent runs Scores from the official leaderboard (10 entries).
- Pass Models scored 9 models we track, plus 1 other entry on its leaderboard.
- Pass Newest score The newest is from September 2026, for Gemini 4 Argon.
- Pass Frontier models 4 of the top 10 have a score, led by GPT-6 Astra (#4).
- Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on nashnovaresearch.github.io: 17 rows for models we track and 1 other entry 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.