Meta was very vague in their announcement for this model, but our take (and backroom conversations) is them updating a bunch of their post-training stack from just rejection sampling and DPO to more in-vogue RL methods (like our Tulu).
Specs
Params70B
LicenseLlama-3.3-Community
Similarity · VAIL
VAIL Fingerprint
0120:01d6:02de:049f:08f7:116b:199c:6266
Explore other models with behavioral similarity to Llama-3.3-70B-Instruct.
Adoption · Hugging Face
RAM @ 30d
Relative Adoption Metric: 3.23×. Strong. Measured at 30 days.
Hugging Face Downloads
308.8K
last 30d
13.6M
all time
HF Likes
3K
Relative Adoption Metric contextualizes downloads against the model's size bucket.
Inference · OpenRouter
Tokens/Day
—
below top 50 this week
Peak Tokens/Day
21.8B
28 Mar 2025
Peak Rank
#3
24 Jan 2025
OpenRouter publishes daily token totals for its ~50 most-served models. Days missing from the chart mean Llama-3.3-70B-Instruct fell below that cutoff — not zero usage. OpenRouter logs usage separately per dated model version and per variant (like ":free"); the hub combines provider variants while keeping separately cataloged releases distinct.
Related Models
