Chartography benchmark shows zoom tool lifts Fable 5 from 29% to 73%
echen · x · 2026-07-28
HelloSurgeAI’s Chartography benchmark tests whether models can read professional charts. In the quoted result, adding a zoom tool lifts Fable 5 from 29% to 73% accuracy and Sonnet 5 from 13% to 44%, suggesting chart-reading is highly sensitive to interface support rather than raw model ability alone.
Related event: Claude's Zoom Tool Significantly Boosts Chart Recognition Accuracy(3 posts)→
More from Research
- Enterprise AI turns document permissions into the real security bottleneck — Frequent-Engine-9920 · 2026-07-28
- Google’s VGGRPO improves video generation with latent 4D geometry rewards — jiqizhixin · 2026-07-28
- Why Kimi-K3 likely avoided fully asynchronous RL — novasarc01 · 2026-07-28
- An open-source MCP council checks every quote word-for-word and abstains when unsure — ilyautov · 2026-07-28
- A Fantastic Idea for Injecting New Knowledge into LLMs Without Changing Weights — theomitsa · 2026-07-28
- Gael Varoquaux says PFNs and LLMs are being combined for table strings — GaelVaroquaux · 2026-07-28