Models / Compare /Gemma 4 31B vs Llama 3.3 70B Instruct
Head to head
Gemma 4 31B vs Llama 3.3 70B Instruct
Compared on live catalog data — pricing synced daily, benchmark scores with provenance.
The short version
- Llama 3.3 70B Instruct is cheaper to run hosted: $0.32/M output vs $0.34/M for Gemma 4 31B (1.1× difference at the cheapest tracked provider).
- Gemma 4 31B takes 262K tokens of context vs 131K for Llama 3.3 70B Instruct.
- Gemma 4 31B has downloadable weights (OPEN) so you can self-host it; we hold no downloadable copy of Llama 3.3 70B Instruct, so it runs through a provider's API.
- Gemma 4 31B scores 1498.1274 on Arena Coding vs 1345.5887 for Llama 3.3 70B Instruct.
- Gemma 4 31B scores 1420.5646 on Arena Creative Writing vs 1286.3107 for Llama 3.3 70B Instruct.
| Spec | ||
|---|---|---|
| Vendor | Meta | |
| Parameters | 32.7B | 70.6B |
| 262K | 131K | |
| Weights | OPEN | OPEN* |
| License | Apache License 2.0 | llama3.3 |
| Modality | text+image+video->text | text->text |
| Released | Mar 11, 2026 | Nov 26, 2024 |
| 20.6 GBest | 44.5 GBest | |
| Cheapest hosted | $0.10 / $0.34 per 1M (OpenRouter) | $0.10 / $0.32 per 1M (OpenRouter) |
| Tracked providers | 22 | 17 |
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 →Llama 3.3 70B Instruct — our take
Llama 3.3 is a 70.6-billion-parameter text model from Meta with a 131,072-token request limit and a licence that permits commercial use with specific conditions. It is a strong instruction-follower with wide hosting choice, though its reasoning scores sit below what its size might suggest.
full breakdown →Need more columns? Use theinteractive comparison to add up to four models.