Llama-3.3-70B-Instruct
by Charm Hyper · family llama · listed Apr 2026TW
Popular open Llama workhorse for multilingual chat, coding, and self-hosting
Current rates USD per 1M tokens
Input$0.607
Output$1.04
Cache read—
Cache write$0.303
Against the market
On the record
- Model id
llama-3.3-70b-instruct- Context window
- 128K tokens
- Max output
- 13K tokens
- Input modalities
- text
- Output modalities
- text
- Reasoning
- no
- Tool calls
- yes
- Open weights
- yes
- Knowledge cutoff
- Dec 2023
- Released
- 2026-04-30
- Last updated
- 2026-07-22
- Provider docs
- hyper.charm.land
Back of the envelope
≈
Also listed at
- Nvidiafree in · free out
- Llamafree in · free out
- GitHub Modelsfree in · free out
- NanoGPT$0.05 in · $0.23 out
- Cortecs$0.089 in · $0.275 out
- Kilo Gateway$0.10 in · $0.32 out
- Meganova$0.10 in · $0.30 out
- Nebius Token Factory$0.13 in · $0.40 out
- LLM Gateway$0.13 in · $0.40 out
- Helicone$0.13 in · $0.39 out
- IO.NET$0.13 in · $0.38 out
- OpenRouter$0.13 in · $0.40 out
- NovitaAI$0.135 in · $0.40 out
- Hugging Face$0.59 in · $0.79 out
- Regolo AI$0.60 in · $2.70 out