Llama 3.1 70B Instruct is 37.5× cheaper per million tokens (blended). Claude Opus 5 has the larger context window (1,000,000 vs 131,072 tokens, 7.6×). Llama 3.1 70B Instruct is open-weights; the other is not.

Claude Opus 5 vs Llama 3.1 70B Instruct

Live catalog fields. Best-in-row is highlighted.

Claude Opus 5

Anthropic · anthropic__claude-opus-5

Llama 3.1 70B Instructopen

Meta · meta-llama__llama-3.1-70b-instruct

Identity

Fieldanthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
ProviderAnthropicMeta
Sluganthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
Statuslivelive
Open weightsNoYes
License
HuggingFacemeta-llama/Meta-Llama-3.1-70B-Instruct

Quality

Fieldanthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
Quality index (0–100)91.5

Cost

Fieldanthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
Input $ / 1M tokens$5.00$0.40
Cached input $ / 1M$0.50
Output $ / 1M tokens$25.00$0.40

Context

Fieldanthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
Max context tokens1,000,000131,072
Max output tokens128,00016,384

Modalities

Fieldanthropic__claude-opus-5meta-llama__llama-3.1-70b-instruct
Inputtext, image, filetext
Outputtexttext