token.app › Compare › Gemma 4 31B (free) vs GPT-5.6 Luna
Gemma 4 31B (free) vs GPT-5.6 Luna
Gemma 4 31B (free) is free to use, against $0.225 per million tokens. GPT-5.6 Luna accepts the larger context window (1.1M).
Head to head
Blended price uses a 3:1 input:output mix
| Gemma 4 31B (free) | GPT-5.6 Luna | |
|---|---|---|
| Provider | OpenAI | |
| Input $/1M | Free | $0.100 |
| Output $/1M | Free | $0.600 |
| Blended $/1M | Free | $0.225 |
| Context window | 262K | 1.1M |
| Max output | 33K | 128K |
| Released | Apr 2026 | Jul 2026 |
| Licence | Open weights | Proprietary |
Choose Gemma 4 31B (free) if…
- Costs less per token — Free vs $0.225 blended
- Open weights — self-hostable
Choose GPT-5.6 Luna if…
- Larger context window (1.1M)
- Newer model (released Jul 2026)
FAQ
Which is better, Gemma 4 31B (free) or GPT-5.6 Luna?
We do not hold independently-run benchmark scores covering both Gemma 4 31B (free) and GPT-5.6 Luna, so we make no quality claim. Their pricing and specifications are compared on this page.
Is Gemma 4 31B (free) cheaper than GPT-5.6 Luna?
Gemma 4 31B (free) is cheaper: Free versus $0.225 per million tokens blended at 3:1 input:output. Input alone: Free vs $0.100. Output alone: Free vs $0.600.
What context windows do Gemma 4 31B (free) and GPT-5.6 Luna have?
Gemma 4 31B (free): 262,144 tokens. GPT-5.6 Luna: 1,050,000 tokens.
Who makes Gemma 4 31B (free) and GPT-5.6 Luna?
Gemma 4 31B (free) is made by Google. GPT-5.6 Luna is made by OpenAI.
Related comparisons