Question 1

What is Llama 3.3 70B Instruct?

Accepted Answer

Llama 3.3 70B Instruct is a model from Meta, released on Dec 6, 2024. The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model is optimized for multilingual dialogue use cases and outperforms many of the available open source and closed chat models on common industry benchmarks. Supported languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai. Model Card.

Question 2

How much does Llama 3.3 70B Instruct cost?

Accepted Answer

On the Respan gateway, input starts at $0.12 per 1M tokens on Novita, and output starts at $0.36 per 1M tokens on Novita, after gateway discounts. Averaged across 8 providers, it costs $0.38 per 1M input tokens and $0.57 per 1M output tokens.

Question 3

Which providers serve Llama 3.3 70B Instruct on Respan?

Accepted Answer

8 providers serve Llama 3.3 70B Instruct: Novita, Inference.net, OpenRouter, Parasail, IO.net, Tera, and 1 more. Tera and Google Vertex AI require your own provider API key (BYOK).

Question 4

What is Llama 3.3 70B Instruct's context window?

Accepted Answer

Llama 3.3 70B Instruct accepts up to 131K tokens of context on its largest provider route.

Question 5

How do I use Llama 3.3 70B Instruct with the Respan gateway?

Accepted Answer

Point any OpenAI-compatible SDK at https://api.respan.ai/api/ with your Respan API key and set the model to "llama-3.3-70b-instruct". To set it up with a coding agent, copy the agent setup prompt at the top of this page.

Question 6

How fast is Llama 3.3 70B Instruct?

Accepted Answer

Over the last 30 days of requests through the Respan gateway, the fastest provider was IO.net at 46 tokens per second, and the lowest response latency was 0.5 seconds on IO.net. Speed varies with request size and provider.

Question 7

Where does Llama 3.3 70B Instruct rank on the Respan Index?

Accepted Answer

Llama 3.3 70B Instruct scores 8.6 out of 100 on the Respan Index v1.1, ranking #137 of 138 models. By category: Coding 0.9, Agents 4.1, Reasoning 8.6, Knowledge & language 2.6, and Human preference 32.9. The Respan Index combines independent and lab-reported benchmark results into one score, using benchmark data collected by llmmetric.com.

Question 8

How does Llama 3.3 70B Instruct score on benchmarks?

Accepted Answer

Llama 3.3 70B Instruct has 9 published benchmark scores, including WeirdML 14.4%, BALROG 23%, and GPQA Diamond 50.5%. Each score links to where it was published; the data is collected by llmmetric.com.

Provider	Context	Input	Output
NovitaSave 10%`novita/meta-llama/llama-3.3-70b-instruct`	12K	$0.12/M	$0.36/M
Inference.net`inference_net/llama-3.3-70b-instruct`	131K	$0.13/M	$0.40/M
OpenRouter`openrouter/meta-llama/llama-3.3-70b-instruct`	131K	$0.22/M	$0.50/M
Parasail`parasail/meta-llama/Llama-3.3-70B-Instruct`	131K	$0.22/M	$0.50/M
Parasail`parasail/parasail-llama-33-70b-fp8`	131K	$0.22/M	$0.50/M
IO.net`ionet/meta-llama/Llama-3.3-70B-Instruct`	128K	$0.67/M	$0.83/M
TeraBYOK only`tera/meta-llama/Llama-3.3-70B-Instruct`	128K	$0.72/M	$0.72/M
Google Vertex AIBYOK only`vertex_ai/meta/llama-3.3-70b-instruct-maas`	131K	$0.72/M	$0.72/M

Llama 3.3 70B Instruct

Providers

Certified where it counts.

Pricing

Cost by volume

Throughput and latency

Output speed

Response latency

Benchmarks

Respan Indexv1.1

Coding1

Agents1

Apps