GPT-6.5 'Bel' may be pricey but could make AI cheaper by training better small models
VraserX · x · 2026-10-30
VraserX speculates that the rumored GPT-6.5 (codename Bel) could be expensive as hell yet still make AI cheaper overall: if it helps train much better small models, the real payoff lands everywhere else rather than in another premium tier. The author frames this as pure speculation and says he's more excited about that distillation spillover than the premium tier itself.
More from Models
- Researcher Finds 600+ Bugs in LLM Kernel Code; Claude Blocks Safety Analysis — pmddomingos · 2026-10-31
- Google trains Gemini with RL on simulated patients; double-blind study shows big psychiatric intake gains — SRSchmidgall · 2026-10-31
- Signal65: Snowflake's CoCo agent finishes workflow 4.6x faster than manual — ryanshrout · 2026-10-31
- Mistral Large 4 matches GLM on coding with 47% fewer tokens, halves preview price — petrusenko_max · 2026-10-31
- Intelligence per dollar fell ~1,000,000x in three years, shows double-exponential chart — banteg · 2026-10-31
- Altman: each math AI proof is worth $10M, and cost will drop 100x within a year — Traditional-Chip8339 · 2026-10-31