Google Gemma 3 27B vs Meta Llama 3.2 3B Instruct
Side by side from live data: prices, limits, who serves each model and how reliably, and how they do on our chess arena.
In short
- Llama 3.2 3B Instruct is 1.2× cheaper at a typical mix of 3 input tokens per output token.
- Both accept prompts up to 131K tokens.
- Gemma 3 27B output costs $0.16 per 1M via DeepInfra, against the list price of $0.35.
- Gemma 3 27B is the newer model (released 2025-03-12).
| Gemma 3 27B | Llama 3.2 3B Instruct | |
|---|---|---|
| Vendor | Meta | |
| Released | 2025-03-12 | 2024-09-25 |
| Input / 1M | $0.08 | $0.05 |
| Output / 1M | $0.35 | $0.33 |
| Cached input / 1M | $0.04 | — |
| Blended / 1M (3:1) | $0.147 | $0.12 |
| Context | 131K | 131K |
| Max output | 118K | 118K |
| Accepts | text, image | text |
| Reasoning | No | No |
| Providers on OpenRouter | 5 | 2 |
| Cheapest provider, output | $0.16 | $0.33 |
| Best provider uptime, 24 h | 99.72% | 100% |
| Chess arena rating | not playing | not playing |
| Vendor API status | Operational | — |
Cost calculator
Gemma 3 27B
$10.05 per month$0.34 per day
Llama 3.2 3B Instruct
$7.95 per month$0.27 per day
Llama 3.2 3B Instruct saves $2.10 a month on this workload.
At list prices per 1M tokens; prompt caching and batch discounts are not included.
Compare Gemma 3 27B with
Compare Llama 3.2 3B Instruct with