The catalog
What we review.
A forecast is priced against this catalog on its run date. Prices and models move constantly — the date is the honest part of the claim.
Providers covered
Anthropic · OpenAI · Google · xAI · DeepSeek · Moonshot · MiniMax
Fourteen models across seven providers. Published list prices below; the model performance metrics that shape a forecast come from independent benchmarks (Artificial Analysis) and are set per engagement, not shown here.
| Model | Provider | Input / 1M | Output / 1M | Date priced |
|---|---|---|---|---|
| Claude Fable 5 | Anthropic | $10.00 | $50.00 | Jul 2026 |
| Claude Opus 5 | Anthropic | $5.00 | $25.00 | Jul 2026 |
| Claude Opus 4.8 | Anthropic | $5.00 | $25.00 | Jul 2026 |
| Claude Sonnet 5 | Anthropic | $3.00 | $15.00 | Jul 2026 |
| Claude Haiku 4.5 | Anthropic | $1.00 | $5.00 | Jul 2026 |
| GPT-5.6 Sol (max) | OpenAI | $5.00 | $30.00 | Jul 2026 |
| GPT-5.6 Luna (max) | OpenAI | $1.00 | $6.00 | Jul 2026 |
| Gemini 3.1 Pro Preview | $2.00 | $12.00 | Jul 2026 | |
| Gemini 3.5 Flash | $1.50 | $9.00 | Jul 2026 | |
| Grok 4.5 (high) | xAI | $2.00 | $6.00 | Jul 2026 |
| DeepSeek V4 Flash (max) | DeepSeek | $0.08 | $0.17 | Jul 2026 |
| Kimi K2.6 | Moonshot | $0.66 | $3.41 | Jul 2026 |
| Kimi K3 | Moonshot | $3.00 | $15.00 | Jul 2026 |
| MiniMax-M3 | MiniMax | $0.30 | $1.20 | Jul 2026 |
Why it's dated
A forecast is a point in time.
Prices and models move constantly. A forecast is priced against the catalog on its run date — and nothing more. Ongoing calibration is how that answer stays current as the leaderboard changes underneath it.
Where the numbers come from
Grounded, not guessed.
Our model performance metrics are grounded by industry benchmarks from Artificial Analysis, then applied to the structure of your workflow. When you reach the calibration tier, your own observed performance replaces the published figures.
Your forecast is priced against this catalog, on the day it runs.
Send us a workflow