VALUE FRONTIER
Performance
per dollar.
Only models with a verified category score and current USD API price appear. Each category stands alone; there is no universal winner.
reasoning evidence vs standard API price
Green points are efficient within this eligible cohort: no other plotted model has both a lower price and an equal or higher score. Prices use the stated 75/25 input/output mix.
| Model | Category score | Blended price | Confidence | Frontier |
|---|---|---|---|---|
| GPT-6 Luna | 39 | $0.20 | Moderate confidence · 4 independent domains | Efficient in this cohort |
| GPT-5.6 Luna | 51 | $0.45 | Moderate confidence · 4 independent domains | Efficient in this cohort |
| DeepSeek V4.1 Flash | 92 | $0.52 | Moderate confidence · 3 independent domains | Efficient in this cohort |
| Gemini 3.5 Flash-Lite | 0 | $0.85 | Limited evidence · 1 independent domains | — |
| Gemini 3.8 Flash | 61 | $1.50 | Moderate confidence · 7 independent domains | — |
| Gemini 3.7 Flash | 91 | $1.50 | Moderate confidence · 4 independent domains | — |
| Gemini 3.6 Flash | 79 | $1.50 | Limited evidence · 2 independent domains | — |
| DeepSeek V4 Pro | 95 | $1.98 | Limited evidence · 2 independent domains | Efficient in this cohort |
| Grok 4.6 | 67 | $3.00 | Moderate confidence · 6 independent domains | — |
| Grok 4.5 | 90 | $3.00 | Moderate confidence · 4 independent domains | — |
| Gemini 3.5 Flash | 75 | $3.38 | Moderate confidence · 4 independent domains | — |
| GPT-6 Sol | 67 | $4.00 | Moderate confidence · 4 independent domains | — |
| GPT-6.1 Sol | 83 | $4.00 | Limited evidence · 4 independent domains | — |
| Claude Sonnet 5.5 | 91 | $4.00 | Moderate confidence · 3 independent domains | — |
| GPT-5.6 Terra | 67 | $4.50 | Moderate confidence · 5 independent domains | — |
| Gemini 3.1 Pro Preview | 81 | $4.50 | Moderate confidence · 4 independent domains | — |
| GPT-5.6 Sol | 69 | $8.00 | Moderate confidence · 5 independent domains | — |
| Claude Opus 5.5 | 91 | $8.00 | Moderate confidence · 5 independent domains | — |
| GPT-6 Astra | 94 | $20.00 | Moderate confidence · 7 independent domains | — |
| Claude Fable 5.1 | 86 | $20.00 | Moderate confidence · 4 independent domains | — |
Median standard API price
2026-09-292026-10-01
- 2026-09-29: 3 USD / 1M · Median of 24 priced models
- 2026-10-01: 3 USD / 1M · Median of 34 priced models
Qualified performance per dollar
2026-10-012026-11-07
- 2026-10-01: 120 · Best qualified overall score per blended USD
- 2026-10-02: 230 · Best qualified overall score per blended USD
- 2026-10-03: 230 · Best qualified overall score per blended USD
- 2026-10-04: 230 · Best qualified overall score per blended USD
- 2026-10-05: 230 · Best qualified overall score per blended USD
- 2026-10-06: 230 · Best qualified overall score per blended USD
- 2026-10-07: 230 · Best qualified overall score per blended USD
- 2026-10-08: 230 · Best qualified overall score per blended USD
- 2026-10-09: 230 · Best qualified overall score per blended USD
- 2026-10-10: 230 · Best qualified overall score per blended USD
- 2026-10-11: 230 · Best qualified overall score per blended USD
- 2026-10-12: 230 · Best qualified overall score per blended USD
- 2026-10-13: 230 · Best qualified overall score per blended USD
- 2026-10-14: 230 · Best qualified overall score per blended USD
- 2026-10-15: 230 · Best qualified overall score per blended USD
- 2026-10-16: 230 · Best qualified overall score per blended USD
- 2026-10-17: 230 · Best qualified overall score per blended USD
- 2026-10-18: 230 · Best qualified overall score per blended USD
- 2026-10-19: 230 · Best qualified overall score per blended USD
- 2026-10-20: 230 · Best qualified overall score per blended USD
- 2026-10-21: 230 · Best qualified overall score per blended USD
- 2026-10-22: 230 · Best qualified overall score per blended USD
- 2026-10-23: 230 · Best qualified overall score per blended USD
- 2026-10-24: 230 · Best qualified overall score per blended USD
- 2026-10-25: 230 · Best qualified overall score per blended USD
- 2026-10-26: 230 · Best qualified overall score per blended USD
- 2026-10-27: 230 · Best qualified overall score per blended USD
- 2026-10-28: 230 · Best qualified overall score per blended USD
- 2026-10-29: 230 · Best qualified overall score per blended USD
- 2026-10-30: 230 · Best qualified overall score per blended USD
- 2026-10-31: 230 · Best qualified overall score per blended USD
- 2026-11-01: 230 · Best qualified overall score per blended USD
- 2026-11-02: 230 · Best qualified overall score per blended USD
- 2026-11-03: 230 · Best qualified overall score per blended USD
- 2026-11-04: 230 · Best qualified overall score per blended USD
- 2026-11-05: 230 · Best qualified overall score per blended USD
- 2026-11-06: 230 · Best qualified overall score per blended USD
- 2026-11-07: 230 · Best qualified overall score per blended USD
ARC-AGI-3 Standard benchmark frontier
2026-09-292026-09-29
- 2026-09-29: 62.713 % · gpt-6-astra · ARC-AGI-3 Semi-Private Standard harness best effort 3 Semi-Private · source ↗
One observation; no trend can be inferred.