Models / Compare /Gemma 4 31B vs Kimi K2.7 Code
Head to head
Gemma 4 31B vs Kimi K2.7 Code
Compared on live catalog data — pricing synced daily, benchmark scores with provenance.
The short version
- Gemma 4 31B is cheaper to run hosted: $0.34/M output vs $3.50/M for Kimi K2.7 Code (10.3× difference at the cheapest tracked provider).
- Gemma 4 31B has downloadable weights (OPEN) so you can self-host it; we hold no downloadable copy of Kimi K2.7 Code, so it runs through a provider's API.
- Kimi K2.7 Code scores 0.0044 on Arena Agent (IPS) vs -0.161 for Gemma 4 31B.
- Kimi K2.7 Code scores 1472.5712 on Arena Code (WebDev) vs 1363.3922 for Gemma 4 31B.
| Spec | ||
|---|---|---|
| Vendor | Moonshot AI | |
| Parameters | 32.7B | 1.1T / 32B |
| 262K | 262K | |
| Weights | OPEN | OPEN* |
| License | Apache License 2.0 | Custom licence |
| Modality | text+image+video->text | text+image->text |
| Released | Mar 11, 2026 | Jun 11, 2026 |
| 20.6 GBest | 667.5 GBest | |
| Cheapest hosted | $0.10 / $0.34 per 1M (OpenRouter) | $0.73 / $3.50 per 1M (OpenRouter) |
| Tracked providers | 22 | 20 |
Gemma 4 31B — our take
Gemma 4 is a 32.7-billion-parameter text, image and video model from Google with a permissive Apache licence. It scores consistently across six Arena categories, with coding as its relative standout, and is available from ten hosts with a wide spread in price and speed.
full breakdown →Kimi K2.7 Code — our take
Moonshot's very large coding model can be downloaded. At over a trillion parameters it is an unusually large model to be downloadable at all, aimed at agentic coding workloads where scale matters.
full breakdown →Need more columns? Use theinteractive comparison to add up to four models.