Benchmark profile

Vals MortgageTax (MortgageTax)

Vals AI benchmark for mortgage and tax document reasoning, including semantic and numerical extraction task views.

Data verified July 18, 2026

How BenchLM shows MortgageTax

BenchLM mirrors the public Vals AI MortgageTax leaderboard captured from https://www.vals.ai/benchmarks/mortgage_tax and updated by Vals on July 18, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

MortgageTax 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.

87 Vals rows3 task viewsprivate datasetTasks: Overall, Semantic Extraction, Numerical ExtractionDisplay only

MortgageTax on Vals AI Vals methodology Vals home

MortgageTax score on MortgageTax — July 18, 2026

BenchLM mirrors the published mortgagetax score view for MortgageTax. Claude Opus 4.7 leads the public snapshot at 70.27% , followed by Claude Sonnet 5 (70.03%) and Claude Opus 4.8 (69.91%). BenchLM does not use these results to rank models overall.

Claude Opus 4.7

Anthropic

anthropic/claude-opus-4-7

70.27%

Overall —

Claude Sonnet 5

Anthropic

anthropic/claude-sonnet-5

70.03%

Overall —

Claude Opus 4.8

Anthropic

anthropic/claude-opus-4-8

69.91%

Overall —

87 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated July 18, 2026

MortgageTax score table (87 models)

Score

Claude Opus 4.7Anthropic

70.27%

Claude Sonnet 5Anthropic

70.03%

Claude Opus 4.8Anthropic

69.91%

Gemini 3.1 Pro PreviewGoogle

69.40%

Gemini 3 Pro PreviewGoogle

69.08%

Claude Fable 5Anthropic

68.92%

Gemini 2.5 ProGoogle

68.92%

GPT-5.5OpenAI

68.76%

Gemini 3 Flash PreviewGoogle

68.72%

Claude 3.7 SonnetAnthropic

68.68%

Claude Opus 4.5Anthropic

68.68%

Claude Opus 4.6 ThinkingAnthropic

68.52%

MiniMax M3MiniMax

68.36%

GPT-5.4OpenAI

68.32%

Qwen3.6 27bAlibaba

68.28%

Gemini 3.5 FlashGoogle

68.12%

Gemini 3.1 Flash Lite PreviewGoogle

68.04%

Qwen3.6 PlusAlibaba

67.97%

Claude Sonnet 4.6Anthropic

67.73%

Claude Opus 4.5 20251101 ThinkingAnthropic

67.69%

Qwen3.5 FlashAlibaba

67.37%

GPT-5.6 TerraOpenAI

67.33%

GPT-5.6 LunaOpenAI

67.29%

GPT-5.6 SolOpenAI

67.29%

Gemini 2.5 Pro Exp 03 25Google

67.17%

GPT-5.2OpenAI

67.13%

GPT-5 MiniOpenAI

66.89%

Claude 3.7 Sonnet 20250219 ThinkingAnthropic

66.85%

Kimi K2.5 ThinkingMoonshot AI

66.53%

Kimi K3Moonshot AI

66.34%

Qwen3.7 PlusAlibaba

66.18%

Muse Spark 1 1Meta

66.14%

GPT-4.1OpenAI

65.94%

Kimi K2.6Moonshot AI

65.82%

O3OpenAI

65.70%

GPT-4.1 MiniOpenAI

65.50%

GPT-5OpenAI

65.45%

O4 MiniOpenAI

64.83%

Claude 3.5 SonnetAnthropic

64.07%

Claude Sonnet 4.5 20250929 ThinkingAnthropic

63.99%

InklingThinkingmachines

63.99%

GPT-5.4 MiniOpenAI

63.51%

Gemini 2.5 Flash Preview 09 2025Google

62.60%

Claude Sonnet 4Anthropic

62.47%

Claude Haiku 4.5 20251001 ThinkingAnthropic

62.16%

Magistral Small 2509Mistral AI

62.12%

Gemini 2.5 Flash Preview 09 2025 ThinkingGoogle

61.92%

Grok 4.5SpaceXAI

61.80%

Gemma 4 31b ItGoogle

61.37%

GPT-5.1OpenAI

61.37%

Mistral Small 2503Mistral AI

61.21%

Claude Opus 4.1Anthropic

61.09%

GPT-4oOpenAI

60.97%

Qwen3.5 Plus ThinkingAlibaba

60.77%

Gemini 2.0 Flash ExpGoogle

60.37%

Gemini 2.0 Flash 001Google

59.66%

Mimo V2.5Xiaomi

59.26%

GPT-5.4 NanoOpenAI

59.10%

Claude Opus 4Anthropic

58.59%

Llama4 Maverick Instruct BasicFireworks AI

58.51%

Gemini 2.5 Flash Preview 04 17Google

58.43%

Meta Llama Llama 4 Scout 17B 16E InstructTogether AI

57.75%

Gemini 2.5 Flash Lite Preview 09 2025 ThinkingGoogle

57.55%

Gemini 2.5 Flash Preview 04 17 ThinkingGoogle

57.51%

GPT-4oOpenAI

57.43%

Gemini 1.5 Pro 002Google

56.76%

Claude Opus 4.1 20250805 ThinkingAnthropic

56.12%

Gemini 2.5 Flash Lite Preview 09 2025Google

55.80%

GPT-4o MiniOpenAI

54.49%

GPT-5 NanoOpenAI

53.62%

GPT-4.1 NanoOpenAI

52.82%

Mistral Large 2512Mistral AI

52.11%

Magistral Medium 2509Mistral AI

51.87%

Claude Sonnet 4 20250514 ThinkingAnthropic

49.88%

Grok 4.3SpaceXAI

48.25%

Grok 4.20 0309 ReasoningSpaceXAI

45.35%

Grok 4 0709SpaceXAI

44.48%

Gemini 1.5 Flash 002Google

42.77%

Grok 4.1 Fast ReasoningSpaceXAI

42.61%

Grok 4 Fast ReasoningSpaceXAI

42.09%

Meta Llama Llama 3.2 90B Vision Instruct TurboTogether AI

38.79%

Mistral Medium 2505Mistral AI

36.45%

Grok 4.1 Fast Non ReasoningSpaceXAI

34.26%

Mistral Medium 3.5Mistral AI

28.89%

Grok 4 Fast Non ReasoningSpaceXAI

25.32%

Meta Llama Llama 3.2 11B Vision Instruct TurboTogether AI

23.29%

Grok 2 Vision 1212SpaceXAI

3.46%

The published MortgageTax snapshot places Claude Opus 4.7 first at 70.27%. The third row is 0.36 points behind. The broader top-10 range is 1.59 points, so many of the published results sit in a relatively narrow band.

87 models have been evaluated on MortgageTax. The benchmark falls in the External benchmark mirrors category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. MortgageTax is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MortgageTax

Year

2026

Tasks

Mortgage and tax extraction tasks

Format

Accuracy score

Difficulty

Professional mortgage-tax document reasoning

BenchLM mirrors Vals MortgageTax as a display-only finance and document-reasoning benchmark.

MortgageTax Public benchmark source

BenchLM freshness & provenance

Version

MortgageTax 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

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.

FAQ

What does MortgageTax measure?

Vals AI benchmark for mortgage and tax document reasoning, including semantic and numerical extraction task views.

Which model leads the published MortgageTax snapshot?

Claude Opus 4.7 currently leads the published MortgageTax snapshot with 70.27% mortgagetax score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on MortgageTax?

87 AI models are included in BenchLM's mirrored MortgageTax snapshot, based on the public leaderboard captured on July 18, 2026.

Last updated: July 18, 2026 · mirrored from the public benchmark leaderboard

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.