Models·Model board

Reasoning Models in 2026

This board covers reasoning models from providers such as OpenAI and SpaceXAI. Use it to compare model names, families, release dates and the provider link before connecting a model to your software.

Read the signal score out of 10 alongside the API, open-source and free-tier marks. For a closer fit, check the exact model variant, including labels such as Pro, batch, Luna or Multi-Agent.

  • 332 models
  • Checked 16 Oct 2026
  1. 1
    S
    Grok 4.20SpaceXAI · reasoning · vision
    $1.25in / 1M $2.50out / 1M 2M context
  2. 2
    S
    Grok 4.20 Multi-AgentSpaceXAI · reasoning · vision
    $1.25in / 1M $2.50out / 1M 2M context
  3. 3
    O
    GPT-5.4OpenAI · reasoning · vision
    $2.50in / 1M $15out / 1M 1.05M context
  4. 4
    O
    GPT-5.4 (batch)OpenAI · reasoning · vision
    $1.25in / 1M $7.50out / 1M 1.05M context
  5. 5
    O
    GPT-5.4 ProOpenAI · reasoning · vision
    $30in / 1M $180out / 1M 1.05M context
  6. 6
    O
    GPT-5.4 Pro (batch)OpenAI · reasoning · vision
    $15in / 1M $90out / 1M 1.05M context
  7. 7
    O
    GPT-5.5OpenAI · reasoning · vision
    $5in / 1M $30out / 1M 1.05M context
  8. 8
    O
    GPT-5.5 (batch)OpenAI · reasoning · vision
    $2.50in / 1M $15out / 1M 1.05M context
  9. 9
    O
    GPT-5.5 ProOpenAI · reasoning · vision
    $30in / 1M $180out / 1M 1.05M context
  10. 10
    O
    GPT-5.5 Pro (batch)OpenAI · reasoning · vision
    $15in / 1M $90out / 1M 1.05M context
  11. 11
    O
    GPT-5.6 LunaOpenAI · reasoning · vision
    $0.20in / 1M $1.20out / 1M 1.05M context
  12. 12
    O
    GPT-5.6 Luna (batch)OpenAI · reasoning · vision
    $0.10in / 1M $0.60out / 1M 1.05M context
  13. 13
    O
    GPT-5.6 Luna ProOpenAI · reasoning · vision
    $0.20in / 1M $1.20out / 1M 1.05M context
  14. 14
    O
    GPT-5.6 Luna Pro (batch)OpenAI · reasoning · vision
    $0.10in / 1M $0.60out / 1M 1.05M context
  15. 15
    O
    GPT-5.6 SolOpenAI · reasoning · vision
    $2in / 1M $10out / 1M 1.05M context
  16. 16
    O
    GPT-5.6 Sol (batch)OpenAI · reasoning · vision
    $1in / 1M $5out / 1M 1.05M context
  17. 17
    O
    GPT-5.6 Sol ProOpenAI · reasoning · vision
    $4in / 1M $20out / 1M 1.05M context
  18. 18
    O
    GPT-5.6 Sol Pro (batch)OpenAI · reasoning · vision
    $1in / 1M $5out / 1M 1.05M context
  19. 19
    O
    GPT-5.6 TerraOpenAI · reasoning · vision
    $2in / 1M $12out / 1M 1.05M context
  20. 20
    O
    GPT-5.6 Terra (batch)OpenAI · reasoning · vision
    $1in / 1M $6out / 1M 1.05M context
  21. 21
    O
    GPT-5.6 Terra ProOpenAI · reasoning · vision
    $2in / 1M $12out / 1M 1.05M context
  22. 22
    O
    GPT-5.6 Terra Pro (batch)OpenAI · reasoning · vision
    $1in / 1M $6out / 1M 1.05M context
  23. 23
    O
    GPT-6.1 SolOpenAI · reasoning · vision
    $2in / 1M $10out / 1M 1.05M context
  24. 24
    O
    GPT-6.1 Sol ProOpenAI · reasoning · vision
    $2in / 1M $10out / 1M 1.05M context
  25. 25
    O
    GPT-6 AstraOpenAI · reasoning · vision
    $10in / 1M $50out / 1M 1.05M context

How to choose

Match the exact variant

Names can differ by plan or workflow: GPT-5.4, GPT-5.4 Pro and GPT-5.4 (batch) are separate entries. Grok 4.20 and Grok 4.20 Multi-Agent are separate models too.

Check the full name in your software settings or API request so you select the intended variant.

Check the provider and family

The list identifies a maker and family for each model. OpenAI entries use the openai family, while Grok entries use x-ai and list SpaceXAI as maker.

Use those labels to keep provider accounts, model permissions and billing aligned with your existing setup.

Consider release dates

Release dates range from March through July 2026 in the entries shown. A date helps you distinguish a newer model name from an earlier variant in the same family.

Record the date with your configuration so a later model change is easy to track.

Confirm access before choosing

Use the API, open-source and free-tier marks to check how a model can be used. A public API suits software connections; open-source code matters when you need to inspect or change code.

Follow the official model link for the exact entry, including batch and Pro variants.

Questions

What does the signal score show?

It is a score out of 10 based on reported use, a public API, documentation depth, open-source code and a free tier, using the stated percentage weights.

Why are batch and Pro names separate?

They are listed as distinct model entries. Check the provider’s description and your software’s integration before selecting a variant.

Which makers appear in these entries?

The entries shown come from OpenAI and SpaceXAI. Their families are listed as openai and x-ai.

How can I identify a model in an API setup?

Use the full model name and its official link. Names such as GPT-5.5 Pro (batch) and Grok 4.20 Multi-Agent identify distinct entries.

Prices are each provider’s published price per million tokens. How we rank · Disclosure