Meta · open weights · small
Meta: Llama 3.1 8B Instruct
Open weights, so you can run it locally. The published id includes 8B. That label is copied from the id, not measured here. Size class is this site's filter bucket, not a benchmark.
API price, in pounds
- Input, per million
- £0.0149
- Output, per million
- £0.0298
OpenRouter lists $0.0200 in and $0.0400 out per million tokens. Converted at 0.74501 on 2026-11-10. Fetched 2026-11-11. Context length listed as 131,072 tokens.
Sourced local speeds
- 160.8 reply tokens/sec on GeForce RTX 4090, Q4_K_M. Source. Run r_32ipo1hbbwl, marked Measured by llm-speed. Chat-short workload. Power draw was not published, so electricity uses the card's official total graphics power and is labelled an estimate. Page read 11 November 2026.
Comparisons
- Meta: Llama 3.1 8B Instruct on GeForce RTX 5090
- Meta: Llama 3.1 8B Instruct on GeForce RTX 3090
- Meta: Llama 3.1 8B Instruct on GeForce RTX 4090
- Meta: Llama 3.1 8B Instruct on GeForce RTX 4070 Ti SUPER
- Meta: Llama 3.1 8B Instruct on GeForce RTX 4080 SUPER
- Meta: Llama 3.1 8B Instruct on GeForce RTX 5070 Ti