Meta · Efficient · Open weights
Llama 3.1 8B
A small open-weight model that runs comfortably on modest hardware. A long-standing default for local experimentation.
Strong at
- -Small footprint
- -Open weights
- -Easy to self-host
Typical use
- -Local prototyping
- -Edge deployment
Same tier
What else to look at
| Model | Developer | Context | In / 1M | Out / 1M |
|---|---|---|---|---|
| Claude Haiku 5.5 | Anthropic | 1M | $0.10 | $0.50 |
| Claude Haiku 4.5 | Anthropic | 200K | $1 | $5 |
| GPT-6 Luna | OpenAI | 1.05M | $0.10 | $0.50 |
| GPT-5.6 Luna | OpenAI | 1.05M | $0.20 | $1.20 |
| Gemini 3.5 Flash-Lite | 1.05M | $0.30 | $2.50 |
Verified 2026-10-11.