Vals Finance Agent
Vals AI / Professional work · Everyday financial-analyst tasks assess work such as research reports and modeling.
Across e/accCounts toward ranking
Evidence budget3%
Upstream data time10/16 08:00
Last successful sync10/17 08:05
What it measures and how
Use only the official Finance Agent v2 Overall accuracy.
How this evidence is used
Receives the 3% industry-specific tasks budget. Currently measures finance directly, without claiming coverage of every industry.
Benchmark results
Fixed rules select one representative configuration per public model. Anonymous test models are omitted. Original source ranks are retained, with up to 30 models per benchmark.
| Source rank | Source model name | Raw result | Representative configuration |
|---|---|---|---|
| 1 | google/gemini-4-argonGoogle | 65.4% | High reasoning |
| 2 | google/gemini-3.8-flashGoogle | 61.4% | High reasoning |
| 3 | meta/muse-spark-openMeta | 61.3% | xHigh reasoning |
| 4 | meta/muse_spark_1_2Meta | 60.6% | xHigh reasoning |
| 5 | meta/muse_spark_1_3_maxMeta | 60.0% | Max reasoning |
| 6 | meta/muse-wonder-fullMeta | 59.1% | Max reasoning |
| 7 | google/gemini-3.7-flashGoogle | 59.0% | High reasoning |
| 9 | anthropic/claude-fable-5-1Anthropic | 58.9% | Max reasoning |
| 10 | anthropic/claude-opus-5Anthropic | 58.6% | Max reasoning |
| 11 | anthropic/claude-opus-5-5Anthropic | 58.6% | Max reasoning |
| 12 | anthropic/claude-sonnet-5-5Anthropic | 58.1% | Max reasoning |
| 13 | google/gemini-3.5-flashGoogle | 57.9% | High reasoning |
| 14 | zai/glm-5.3-flashZ.ai | 57.9% | Max reasoning |
| 15 | xiaomi/mimo-v2.6-proXiaomi | 57.3% | Source default configuration |
| 16 | meta/muse_spark_1_1Meta | 57.2% | xHigh reasoning |
| 17 | meta/muse-wonderMeta | 56.6% | Max reasoning |
| 18 | anthropic/claude-fable-5Anthropic | 56.3% | Max reasoning |
| 19 | google/gemini-3.6-flashGoogle | 56.3% | High reasoning |
| 20 | xiaomi/mimo-v2.6-flashXiaomi | 56.3% | Source default configuration |
| 21 | zai/glm-5.3Z.ai | 55.8% | Max reasoning |
| 22 | tencent/hy4-previewTencent | 55.1% | Source default configuration |
| 23 | openai/gpt-5.6-lunaOpenAI | 55.0% | Max reasoning |
| 24 | ant/ling-3.0-flash-af-rc3Ant | 54.9% | Source default configuration |
| 25 | mistralai/mistral-large-4Mistral AI | 54.7% | High reasoning |
| 26 | openai/gpt-5.6-terraOpenAI | 54.4% | Max reasoning |
| 27 | anthropic/claude-haiku-5-5Anthropic | 54.1% | Max reasoning |
| 28 | anthropic/claude-opus-4-8Anthropic | 53.9% | Max reasoning |
| 29 | anthropic/claude-sonnet-5Anthropic | 53.9% | Max reasoning |
| 30 | openai/gpt-5.6-solOpenAI | 53.8% | Max reasoning |
| 31 | grok/grok-4.6xAI | 53.7% | High reasoning |
Limitations and attribution
Private tasks prevent task-by-task replication. Like APEX, these are office tasks and must share a budget group.
Data license: Official public results; public display retains official attribution. Full redistribution authorization remains subject to Vals terms.
Results are published by Vals AI. Raw scores and the e/acc consensus index use different scales and cannot be added directly.