Vals Finance Agent v1.1 (Finance Agent v1.1)
We show this table for reference; we do not rank on it.
An archived Vals benchmark covering retrieval, numerical reasoning, financial modeling, market analysis, earnings, trends, and adjustments.
Evaluation results published by Vals AI. Published leaderboard aggregates and evaluation settings only. Dataset contents are not included.
Finance Agent v1.1 score on Finance Agent v1.1 — June 4, 2026
We mirror the published finance agent v1.1 score view for Finance Agent v1.1. Claude Opus 4.7 leads the public snapshot at 64.37%, followed by Claude Sonnet 4.6 (63.33%) and Muse Spark (60.59%). We do not use these results to rank models overall.
Claude Opus 4.7
Anthropic
Claude Sonnet 4.6
Anthropic
Muse Spark
Meta
51 modelsAgenticCurrentDisplay onlyUpdated June 4, 2026
Finance Agent v1.1 score table (51 models)
| Rank | Model / configuration | Score | Parameters (B) | Open / closed |
|---|---|---|---|---|
| 1 | Claude Opus 4.7Anthropic | 64.37% | Not reported | Closed |
| 2 | Claude Sonnet 4.6Anthropic | 63.33% | Not reported | Closed |
| 3 | Muse SparkMeta | 60.59% | Not reported | Closed |
| 4 | DeepSeek V4 Pro 0813DeepSeekmax reasoning | 60.39% | Not reported | Open |
| 5 | Claude Opus 4.6 (Adaptive)Anthropic | 60.05% | Not reported | Closed |
| 6 | GPT-5.5OpenAIxhigh reasoning | 59.96% | Not reported | Closed |
| 7 | Gemini 3.1 Pro PreviewGooglehigh reasoning | 59.72% | Not reported | Unknown |
| 8 | Claude Opus 4.5 ThinkingAnthropic | 58.81% | Not reported | Closed |
| 9 | GPT-5.2OpenAIxhigh reasoning | 58.53% | Not reported | Closed |
| 10 | GLM-5.1Z.AI | 57.66% | Not reported | Open |
| 11 | GPT-5.4OpenAIxhigh reasoning | 57.15% | Not reported | Closed |
| 12 | Kimi K2.6Moonshot AI | 57.06% | Not reported | Open |
| 13 | GPT-5.1OpenAIhigh reasoning | 55.31% | Not reported | Closed |
| 14 | Gemini 3 Pro PreviewGooglehigh reasoning | 55.15% | Not reported | Unknown |
| 15 | Qwen3.6 PlusAlibaba | 54.63% | Not reported | Closed |
| 16 | Claude Sonnet 4.5 ThinkingAnthropic | 54.50% | Not reported | Closed |
| 17 | Qwen3.5 Plus ThinkingAlibaba | 54.48% | Not reported | Unknown |
| 18 | Grok 4.3xAI | 53.81% | Not reported | Closed |
| 19 | Grok 4 0709xAI | 53.51% | Not reported | Unknown |
| 20 | GPT-5.4 miniOpenAIxhigh reasoning | 53.41% | Not reported | Closed |
| 21 | GLM 5 ThinkingZhipu AI | 53.18% | Not reported | Unknown |
| 22 | Qwen 3.6 Max (preview)Alibaba | 52.78% | Not reported | Closed |
| 23 | Grok 4.1 Fast (Reasoning)xAI | 52.45% | Not reported | Closed |
| 24 | Grok 4.20 0309 ReasoningxAI | 52.30% | Not reported | Unknown |
| 25 | GPT-5OpenAIhigh reasoning | 52.15% | Not reported | Unknown |
| 26 | GPT-5 miniOpenAIhigh reasoning | 51.93% | Not reported | Closed |
| 27 | Gemma 4 31b ItGooglehigh reasoning | 50.79% | Not reported | Unknown |
| 28 | Kimi K2.5 ThinkingMoonshot AI | 50.62% | Not reported | Unknown |
| 29 | MiniMax M2.7MiniMax | 48.40% | Not reported | Open |
| 30 | GPT-5.4 nanoOpenAIhigh reasoning | 47.80% | Not reported | Closed |
| 31 | Gemini 3 Flash PreviewGooglehigh reasoning | 47.60% | Not reported | Unknown |
| 32 | Claude Haiku 4.5 ThinkingAnthropic | 46.93% | Not reported | Closed |
| 33 | Gemini 3.1 Flash Lite PreviewGooglehigh reasoning | 46.12% | Not reported | Unknown |
| 34 | Mistral Medium 3.5Mistral AIhigh reasoning | 46.11% | Not reported | Unknown |
| 35 | Grok 4 Fast (Reasoning)xAI | 46.08% | Not reported | Closed |
| 36 | GLM-4.7Z.AI | 45.98% | Not reported | Open |
| 37 | Qwen3.5 FlashAlibaba | 45.64% | Not reported | Closed |
| 38 | Grok 4.1 Fast Non ReasoningxAI | 44.36% | Not reported | Unknown |
| 39 | Qwen3 MaxAlibaba | 44.30% | Not reported | Closed |
| 40 | Gemini 2.5 ProGoogle | 41.59% | Not reported | Closed |
| 41 | MiniMax M2.5MiniMax | 38.58% | Not reported | Closed |
| 42 | Kimi K2 ThinkingMoonshot AI | 36.65% | Not reported | Unknown |
| 43 | GLM-4.6Z.AI | 36.48% | Not reported | Open |
| 44 | MiniMax M2.1MiniMax | 33.35% | Not reported | Unknown |
| 45 | GPT-OSS 120BOpenAI | 21.54% | Not reported | Open |
| 46 | Mistral Large 2512Mistral AI | 18.05% | Not reported | Unknown |
| 47 | GPT-4oOpenAIhigh reasoning | 8.06% | Not reported | Closed |
| 48 | Command A 03 2025Cohere | 4.23% | Not reported | Unknown |
| 49 | DeepSeek V3p2 ThinkingFireworks AIhigh reasoning | 2.35% | Not reported | Unknown |
| 50 | Jamba Large 1.7AI21 Labs | 0.37% | Not reported | Unknown |
| 51 | DeepSeek V3p2Fireworks AInone reasoning | 0.00% | Not reported | Unknown |
How Finance Agent v1.1 is shown here
BenchLM mirrors the public Vals AI Finance Agent v1.1 leaderboard captured from https://www.vals.ai/benchmarks/finance_agent and updated by Vals on June 4, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.
Finance Agent v1.1 is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.
Snapshot
The published Finance Agent v1.1 snapshot places Claude Opus 4.7 first at 64.37%. The third row is 3.78 points behind. The broader top-10 range is 6.72 points, so many of the published results sit in a relatively narrow band.
51 models have been evaluated on Finance Agent v1.1. The benchmark falls in the Agentic category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Finance Agent v1.1 is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.
About Finance Agent v1.1
Year
2026
Tasks
11 financial analyst task views
Format
Accuracy with task-level breakdowns
Difficulty
Professional financial analysis
We mirror all 51 published rows and 11 task views for historical comparison. Finance Agent v2 is the current Vals benchmark, so v1.1 stays display only and does not enter weighted rankings.
Freshness and provenance
Version
Finance Agent v1.1 2026
Refresh cadence
Quarterly
Staleness state
Current
Question availability
Public benchmark set
BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.
Questions
What does Finance Agent v1.1 measure?
An archived Vals benchmark covering retrieval, numerical reasoning, financial modeling, market analysis, earnings, trends, and adjustments.
Which model leads the published Finance Agent v1.1 snapshot?
Claude Opus 4.7 currently leads the published Finance Agent v1.1 snapshot with 64.37% finance agent v1.1 score. BenchLM shows this benchmark for display only and does not use it in overall rankings.
How many models are evaluated on Finance Agent v1.1?
The June 4, 2026 snapshot contains 51 AI models.
Know when it’s worth switching models
The model to choose, the cheaper alternative, and the release we would wait on.
Read a sample issueJoin 5,600+ readers.
One email each week. Unsubscribe anytime.