MortgageTax (Vals AI) leaderboard
- Claude Opus 5 leads
- 72.1%
- models measured
- 68
- lead over #2
- +1.3 pts
- median
- 65.8%
Leader on MortgageTax (Vals AI) · read October 22
On October 22, 2026, 71 of our catalog's models had a MortgageTax (Vals AI) result, and Anthropic Claude Opus 5 led with 72.1%, ahead of Anthropic Claude Fable 5.1 at 70.8%.
Reading tax certificates from scanned images. Share answered correctly, as run by Vals AI. An archived board.
Role in the index
MortgageTax (Vals AI) is shown as published and does not feed any of our indexes. Read the methodology
| # | Model | Result | Latency, as printed | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 5Anthropic | 72.1% | 8.0 s | 83.0 | $25.00 | |
| 2 | Claude Fable 5.1Anthropic | 70.8% | 15 s | 84.2 | $50.00 | |
| 3 | Claude Opus 4.7Anthropic | 70.3% | 16 s | 67.5 | $25.00 | |
| 4 | Claude Sonnet 5Anthropic | 70.0% | 29 s | 57.7 | $10.00 | |
| 5 | Claude Opus 4.8Anthropic | 69.9% | 21 s | 65.0 | $25.00 | |
| 6 | Gemini 3.1 Pro PreviewGoogle | 69.4% | 24 s | 32.3 | $12.00 | |
| 7 | Claude Fable 5Anthropic | 68.9% | 16 s | 79.3 | $50.00 | |
| 8 | Gemini 2.5 ProGoogle | 68.9% | 24 s | 19.1 | $10.00 | |
| 9 | GPT-5.5OpenAI | 68.8% | 28 s | 71.3 | $30.00 | |
| 10 | Gemini 3 Flash PreviewGoogle | 68.7% | 13 s | 43.7 | $3.00 | |
| 11 | Claude Opus 4.5Anthropic | 68.7% | 11 s | 44.0 | $25.00 | |
| 12 | Gemini 3.5 Flash LiteGoogle | 68.7% | 4.9 s | 49.0 | $2.50 | |
| 13 | Claude Opus 4.6Anthropic | 68.5% | 27 s | 57.2 | $25.00 | |
| 14 | MiniMax M3MiniMax | 68.4% | 27 s | 39.2 | $1.20 | |
| 15 | GPT-5.4OpenAI | 68.3% | 45 s | 64.4 | $15.00 | |
| 16 | Qwen3.6 27BQwen | 68.3% | 42 s | 36.8 | $2.70 | |
| 17 | Gemini 3.5 FlashGoogle | 68.1% | 11 s | 39.8 | $9.00 | |
| 18 | Gemini 3.1 Flash Lite PreviewGoogle | 68.0% | 5.4 s | 22.7 | $1.50 | |
| 19 | Qwen3.6 PlusQwen | 68.0% | 1m 18s | 44.1 | $1.95 | |
| 20 | Claude Sonnet 4.6Anthropic | 67.7% | 30 s | 57.3 | $15.00 | |
| 21 | Gemini 3.6 FlashGoogle | 67.6% | 9.2 s | 62.5 | $3.75 | |
| 22 | Qwen3.5-FlashQwen | 67.4% | 37 s | 41.1 | $0.26 | |
| 23 | GPT-5.6 TerraOpenAI | 67.3% | 6.3 s | 63.9 | $12.00 | |
| 24 | GPT-5.6 LunaOpenAI | 67.3% | 23 s | 58.0 | $1.20 | |
| 25 | GPT-5.6 SolOpenAI | 67.3% | 13 s | 73.5 | $20.00 | |
| 26 | GPT-5.2OpenAI | 67.1% | 1m 11s | 55.9 | $14.00 | |
| 27 | GPT-5 MiniOpenAI | 66.9% | 25 s | 34.2 | $2.00 | |
| 28 | Gemini 3.7 FlashGoogle | 66.7% | 4.1 s | 58.0 | $3.75 | |
| 29 | Kimi K2.5Moonshot AI | 66.5% | 1m 20s | 38.6 | $2.85 | |
| 30 | Kimi K3Moonshot AI | 66.3% | 1m 19s | 62.3 | $15.00 | |
| 31 | Qwen3.7 PlusQwen | 66.2% | 1m 01s | 33.2 | $1.28 | |
| 32 | Muse Spark 1.1Meta | 66.1% | 25 s | 70.3 | $4.25 | |
| 33 | GPT-4.1OpenAI | 65.9% | 5.0 s | Not benchmarked | $8.00 | |
| 34 | Kimi K2.6Moonshot AI | 65.8% | 1m 50s | 52.1 | $4.00 | |
| 35 | o3OpenAI | 65.7% | 21 s | Not benchmarked | $8.00 | |
| 36 | GPT-4.1 MiniOpenAI | 65.5% | 4.6 s | Not benchmarked | $1.60 | |
| 37 | GPT-5OpenAI | 65.5% | 1m 03s | 38.8 | $10.00 | |
| 38 | Muse Spark 1.2Meta | 65.4% | 45 s | 55.7 | $4.25 | |
| 39 | Gemini 3.8 FlashGoogle | 65.3% | 15 s | 61.6 | $3.75 | |
| 40 | Qwen3.8 27BQwen | 64.9% | 1m 20s | 74.2 | $2.55 | |
| 41 | o4 MiniOpenAI | 64.8% | 13 s | Not benchmarked | $4.40 | |
| 42 | Grok 4.6xAI | 64.2% | 8.6 s | 65.7 | $6.00 | |
| 43 | Claude Sonnet 4.5Anthropic | 64.0% | 48 s | 39.5 | $15.00 | |
| 44 | Qwen3.8 Max (0902)Qwen | 64.0% | 48 s | 62.4 | $6.00 | |
| 45 | T InklingThinking Machines | 64.0% | 1m 09s | 31.5 | $4.05 | |
| 46 | GPT-5.4 MiniOpenAI | 63.5% | 2m 06s | 35.8 | $4.50 | |
| 47 | T Inkling SmallThinking Machines | 62.3% | 1m 52s | 27.3 | $1.20 | |
| 48 | Claude Haiku 4.5Anthropic | 62.2% | 30 s | 35.6 | $5.00 | |
| 49 | Grok 4.5xAI | 61.8% | 1m 57s | 58.6 | $6.00 | |
| 50 | Gemma 4 31BGoogle | 61.4% | 42 s | 39.0 | $0.40 | |
| 51 | GPT-5.1OpenAI | 61.4% | 46 s | 40.4 | $10.00 | |
| 52 | Claude Opus 4.1Anthropic | 61.1% | 10 s | Not benchmarked | $75.00 | |
| 53 | GPT-4o (2024-08-06)OpenAI | 61.0% | 7.4 s | Not benchmarked | $10.00 | |
| 54 | Qwen3.5 Plus 2026-04-20Qwen | 60.8% | 1m 02s | 39.5 | $1.80 | |
| 55 | Gemini 2.0 FlashRetiredGoogle | 59.7% | 3.8 s | Not benchmarked | $0.40 | |
| 56 | MiMo-V2.5Xiaomi | 59.3% | 33 s | 50.0 | $0.336 | |
| 57 | GPT-5.4 NanoOpenAI | 59.1% | 15 s | 44.8 | $1.25 | |
| 58 | Llama 4 MaverickMeta Llama | 58.5% | 2.7 s | Not benchmarked | $1.00 | |
| 59 | Llama 4 ScoutMeta Llama | 57.8% | 2.4 s | Not benchmarked | $0.50 | |
| 60 | Gemini 2.5 Flash LiteGoogle | 57.6% | 28 s | Not benchmarked | $0.40 | |
| 61 | GPT-4o (2024-11-20)OpenAI | 57.4% | 6.0 s | Not benchmarked | $10.00 | |
| 62 | Gemini 1.5 ProRetiredGoogle | 56.8% | 4.2 s | Not benchmarked | $5.00 | |
| 63 | GPT-4o-miniOpenAI | 54.5% | 6.1 s | Not benchmarked | $0.60 | |
| 64 | GPT-5 NanoOpenAI | 53.6% | 33 s | Not benchmarked | $0.40 | |
| 65 | GPT-4.1 NanoOpenAI | 52.8% | 3.9 s | Not benchmarked | $0.40 | |
| 66 | Mistral Large 3 2512Mistral AI | 52.1% | 9.3 s | Not benchmarked | $1.50 | |
| 67 | Grok 4.3xAI | 48.3% | 7m 23s | 37.4 | $2.50 | |
| 68 | Grok 4.20xAI | 45.4% | 11 s | 33.3 | $2.50 | |
| 69 | Gemini 1.5 FlashRetiredGoogle | 42.8% | 2.1 s | Not benchmarked | $0.30 | |
| 70 | Mistral Medium 3.1Mistral AI | 36.5% | 7.1 s | Not benchmarked | $2.00 | |
| 71 | Mistral Medium 3.5Mistral AI | 28.9% | 1m 08s | 25.5 | $7.50 |
Only models in our catalog are listed, and each appears once with its best published run. 137 rows were read from the source on October 22, 2026, and rows that match no catalog model are left out. Every other column is a figure the board publishes for the same run.