GPT-6.2-Astra cuts healthcare computer-use agent costs 4.9x in Prescience evals
RishabJainK · x · 2026-11-22
Prescience's internal computer-use evals on real healthcare patient tasks (A/B tested, 24 runs, ABBA order) found GPT-6.2-Astra dramatically cheaper than GPT-6-Astra: mean calls per run dropped 68→40 (1.71×), context per call shrank 51k→29k tokens (1.76×), and tokens are 1.63× cheaper — about 4.9× total. Cost per run fell from $4.44 to $0.91, with per-call pricing at $0.023 vs $0.065. The author notes the model is 'extremely disappointing in other ways.'
More from Models
- Inkling-2 Mini adapts its thinking depth to task size, runs locally at 91 tok/s on RTX 3090 — LiTianleli · 2026-11-22
- She churned through a week's Claude usage researching the Theranos movie — DavidWells · 2026-11-22
- Custom Rust engine runs 524k context on one RX 7900 XTX, 5x faster than llama.cpp — W61k3r · 2026-11-22
- Fine-tune gotcha: HF export silently drops Qwen3's MTP head, slowing decode — AIFlow_ML · 2026-11-22
- Designer KOL: Google's model pace and quality suggest it will soon 'mog them all' — AIandDesign · 2026-11-22
- Models play a literary 'listening at doors' game — Opus 4.5 gets recognized by its timbre — repligate · 2026-11-22