Skip to content
TrustList

TrustList summary

Meta pretrained and instruction tuned Llama 3.3 70B as a single multilingual text model of 70 billion parameters. Its 70B Instruct version was released on December 6, 2024, and Meta describes it as optimized for multilingual dialogue.

Technical profile

Architecturally it is an auto-regressive transformer, and the tuned model gets its helpfulness and safety alignment from supervised fine-tuning and then reinforcement learning from human feedback. It takes multilingual text as input and produces multilingual text and code, with a 128k context length, grouped-query attention, a pretraining set of more than 15 trillion tokens from a new mix of public online data and a December 2023 knowledge cutoff.

Languages, uses and limits

Eight languages are supported: English, German, French, Italian, Portuguese, Hindi, Spanish and Thai, though training covered more and developers can fine-tune for others if they follow the licence and acceptable use policy. Meta intends it for commercial and research use, with the instruction-tuned model suited to assistant-like chat and pretrained models adaptable to other generation tasks. Using its outputs to improve other models, through synthetic data or distillation, is allowed.

Licence

It is distributed under the Llama 3.3 Community License Agreement, a custom commercial licence. Meta notes the training set was offline and fixed, and says tuned versions may follow as safety feedback comes in. The card also reports 39.3 million GPU hours of training compute on H100 hardware.

Summary written by TrustList from the business's own published information, Nov 2026.

Benchmarks & AI stats

BenchmarkOfficialCommunity avg
No benchmark scores yet. Be the first to add one.

“Official” values are editor-approved and feed the ranking. “Community avg” is the mean of member submissions (shown for transparency; it never affects the ranking until an editor approves a value).

Sign in to add a benchmark score for this model.

Llama 3.3 70B launches

Request a demo or quote from Llama 3.3 70B

Protected by reCAPTCHA - Google Privacy Policy and Terms apply.

By sending, you agree we may share your request and contact details with the provider once you confirm your email.

Reviews

Write the first review of Llama 3.3 70B

Used it? Your experience helps other buyers decide.

Write a review

Questions & answers

No questions yet. Be the first to ask about Llama 3.3 70B.

Llama 3.3 70B alternatives

Similar listings buyers compare against this one.

Liquid AI's open-weight decision model, giving yes or no, choice or score answers from text and images

Anthropic's smallest and fastest Claude 5.5 model, available on its own API and three cloud platforms

JetBrains' open 12B mixture-of-experts coding model with 2.5B parameters active per token under Apache 2.0

Cantina Security's first open-weight security model, a post-trained GLM-5.3-Flash for vulnerability research and exploit checks

Microsoft AI's real-time speech-to-text model for voice agents, transcribing 60 languages with continuous language detection

Open-weight translation model from Bilibili's Index team, covering 150 languages and following glossary and style instructions

Aleph Alpha's open-weight English-German mixture-of-experts model for self-hosting, released under Apache 2.0 in October 2026

Perplexity's open-weight decision model that picks among given options for routing and checks in AI systems