VLMs May Turn Clearly Superhuman in Vision Within a Year or Two, Researcher Predicts
arnabdotorg · x · 2026-10-15
Andrew Perrault says the pace of VLM improvement has been shocking — despite the field's focus on coding and RSI, vision progress hasn't slowed, and VLMs will likely be clearly superhuman in both performance and cost within a year or so. His thread splits tasks into acceptability checks and information extraction, notes the best models are already price/performance competitive with a single human rater, cites models' structural advantage of superhuman attention, and shows task-specific routing pays off: on VIN reading, GPT-6 Astra gets 46% right while Sonnet 5 hits 86%.
More from Models
- New SWE-bench Multimodal leaderboard: Claude Opus 5.5 leads at 70.21% on visual bugs — OfirPress · 2026-10-15
- Google says Gemini 4 Argon is delivering breakthroughs in quantum computing, fusion and protein design — GoogleDeepMind · 2026-10-15
- Google sits down with Sergey Brin to discuss Gemini 4 Argon and the road ahead — OfficialLoganK · 2026-10-15
- ChatGPT quietly makes GPT-6.1 Sol the default, users report sharper voice chats — shaunralston · 2026-10-15
- The small-model world has changed forever, but the builder community hasn't priced it in — tokenbender · 2026-10-15
- Gemini 4 Argon tackles real CAD: 263 coding steps to build an actuator base, scores 0.459 — DynamicWebPaige · 2026-10-15