Currently listed through these providers:
Model details
Gemma 4 31B Thinking TEE
Gemma 4 31B Thinking TEE appears in directory listings as an open Gemma-family instruction model oriented around reasoning, surfaced through the models.dev aggregator and attributed to the "Tee" provider brand. Its model identifier carries an explicit ":thinking" tag and the directory entry tags it for reasoning, signaling that the variant is intended to expose visible chain-of-thought style responses for chat and agent-style use cases rather than functioning as a base completion model.
The same directory framing positions this release for efficient chat and self-hosted deployments, aligning it with teams that want a moderately sized, open-weights Gemma variant they can run on their own infrastructure while still benefiting from a reasoning-tuned post-training pass. Because the public evidence for this model is limited to aggregator pages, its strongest practical story is that of a reasoning-capable, self-host-friendly Gemma instruction model, with deeper architectural or benchmark details left for the upstream provider to document.
Quick Info
Powered by- Provider
- NanoGPT
- Model key
- TEE/gemma4-31b:thinking
- Release date
- May 2, 2026
- Last updated
- May 2, 2026
- Input modalities
- Output modalities
- Capabilities
Cost
- Input token cost
- $0.40
- Output token cost
- $1.00
Limits
- Input tokens
- 262,144 tokens
- Output tokens
- 131,072 tokens
- Context window
- 262,144 tokens
Latest news about Gemma 4 31B Thinking TEE
No articles yet. Fetch the latest news to show it here.