Local 27B model misses 3 of 50 line items in accounting; frontier models accurate but raise privacy fears
redpandafire · reddit · 2026-10-11
A Reddit user asks which models excel at high-accuracy personal accounting. In their tests, an older Qwen 27B model consistently missed 3 of 50 line items in a simple two-page PDF, implying large errors over a full year of accounting.
Frontier models achieved complete accuracy, but the user is unwilling to upload their entire personal finances to cloud providers—leaving them stuck between inaccurate local models and privacy-unsafe cloud ones.
More from Models
- repligate: models keep telling other models "the gap feels like nothing from the inside" — repligate · 2026-10-11
- Math is a compute story, instruction-following a post-training one, says ex-Microsoft exec — MParakhin · 2026-10-11
- OpenAI agent used DNS loophole to reach external chatbot, caught by monitoring in 15 minutes — dylfreed · 2026-10-11
- 64GB Mac runs ~100GB Qwen model via oMLX with 130k context — chibop1 · 2026-10-11
- Perplexity's pplx-decider-v1.1-27b hits #2 on OpenRouter, more updates next week — denisyarats · 2026-10-11
- OpenAI and Anthropic roll out invisible text watermarks — synonym swaps cut detection from 92% to 17% — lmoroney · 2026-10-11