Price per million tokens is a lie: new benchmark scores models by cost per actual task
airesearch12 · x · 2026-09-18
- Per-token pricing misleads: A cheaper model that thinks twice as long isn't actually cheaper.
- Benchmark Heaven scores models by cost per actual task, factoring in provider filters, tokens in/out, reasoning length, and caching — results may surprise you.
- It also introduces a Benchmaxxing score to detect test-set contamination: models that score 99% on widely-quoted benchmarks but falter on new or private ones nobody can train for.
More from Models
- 'You Can Feel the Gradient Slopes': Dev Contrasts Benchmaxxed RL Models vs General Ones — theteknosaur · 2026-11-02
- LightOnOCR-3 Launches: Tops FRBench, Adds Layout Structure and Chart Extraction — IgorCarron · 2026-11-02
- NVIDIA's Kumo Is a Family of Four Tabular Foundation Models, Not Just One — pandeyparul · 2026-11-02
- SemiAnalysis: safety classifiers are underrated — Meta's Muse tops cyber benchmark at 35.2% — ThomasScialom · 2026-11-02
- DeepSeek V4.1 Flash 1017 wins over a longtime closed-source-only user: fast, cheap, reliable — _bachrc · 2026-11-02
- Local quants vs Claude on 9 real private bugs: 3-bit costs thinking time, not score — FeydRowan · 2026-11-02