Sources & verification notes

Gemma 4 31B IT

• Exact identifier & access

  Identity and lifecycle apply to this exact identifier and serving surface.

  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• First availability

  Effective Apr 2, 2026. Checked Sep 13, 2026. Re-verification needed.

  • Release notes ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• Availability status

  Effective date not established; this is an observation. Checked Sep 13, 2026. Re-verification needed.

  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• Capabilities & limits

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  Context is the model's 256K-token length (config max_position_embeddings 262144); Google does not publish a separate hosted limit, a maximum output length, or structured-output support for Gemma on the Gemini API.

  • Gemma 4 model card ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • google/gemma-4-31B-it config.json ↗Google (Hugging Face) · provider-authored · checked Sep 23, 2026
  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• Input price

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  Derived from the captured rate card's standard token row.

  • Gemini Developer API pricing ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• Output price

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  Derived from the captured rate card's standard token row.

  • Gemini Developer API pricing ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• Model card

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  • Gemma 4 model card ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • google/gemma-4-31B-it model card ↗Google (Hugging Face) · provider-authored · checked Sep 23, 2026

• Official announcement

  Effective Apr 2, 2026. Checked Sep 23, 2026.

  • Gemma 4: Byte for byte, the most capable open models ↗Google · provider-authored · checked Sep 23, 2026

• Full rate card

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  Google serves Gemma 4 on the Gemini API free of charge on the free tier; the paid tier lists it as not available.

  • Gemini Developer API pricing ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• License

  Effective date not established; this is an observation. Checked Sep 13, 2026. Re-verification needed.

  The hosted endpoint guide identifies Gemma as an open model; exact license text is governed separately by the Gemma Terms.

  • Run Gemma with the Gemini API ↗Google AI for Developers · provider-authored · checked Sep 23, 2026

• License

  Effective date not established; this is an observation. Checked Sep 23, 2026.

  Gemma 4 is released under the Apache License 2.0, not the earlier Gemma Terms of Use.

  • Gemma 4 model card ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • Gemma 4 license (Apache License 2.0) ↗Google AI for Developers · provider-authored · checked Sep 23, 2026
  • google/gemma-4-31B-it model card ↗Google (Hugging Face) · provider-authored · checked Sep 23, 2026
  • Gemma 4: Byte for byte, the most capable open models ↗Google · provider-authored · checked Sep 23, 2026