Llama 3 8B Instruct
Meta's Llama 3 model is rolled out in varying sizes, with the 8B instruct-tuned variant particularly optimized for high-quality dialogue purposes. The performance of Llama 3 has been found to be strong when compared to leading closed-source models, backed by human evaluations. The model family includes 8B and 70B sizes, with pretrained and instruct-tuned variants suitable for assistant-like chat. Llama 3 leverages auto-regressive language modeling with optimized transformer architecture. Training data is a mix of publicly available online data, with a knowledge cutoff in March 2023 for 8B and December 2023 for 70B models. The models use Grouped-Query Attention for improved inference scalability.
Model card on Hugging Face- Input / output price
- $0.30 / $0.60 per 1M
- Context
- 8K tokens
- Released
- Apr 18, 2024
Paste into Claude Code, Cursor, or Codex
Help me set up the Respan gateway: fetch and follow https:/
Providers
2 providers, cheapest first, priced per 1M tokens after gateway discounts.
| Provider | Input | Output |
|---|---|---|
Bedrock bedrock/llama-3-8b-instruct | $0.30/M | $0.60/M |
Concentrate concentrate/llama-3-8b-instruct | $0.30/M | $0.60/M |
Is a provider down? AWS Bedrock status
Pricing
Averaged across 2 providers, in USD per 1M tokens after gateway discounts.
- Average input
- $0.30 per 1M
- Average output
- $0.60 per 1M
Cost by volume
Total cost from 0 to 1T tokens.
Throughput and latency
30-day averages from requests through the Respan gateway. Under each figure, every model on the gateway by its average, this one in blue.
Output speed
Higher is better22tokens/sacross providers
Response latency
Lower is better1.0sacross providers
Benchmarks
5 published scores from 2 sources, each linked to where it was published.
Published scores
| Benchmark | Score | Source | Published |
|---|---|---|---|
| HumanEvalhuggingface.co | 62.2% | huggingface.co | Apr 18, 2024 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| GSM8Khuggingface.co | 79.6% | huggingface.co | Apr 18, 2024 |
| MATHhuggingface.co | 30% | huggingface.co | Apr 18, 2024 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| MMLUhuggingface.co | 68.4% | huggingface.co | Apr 18, 2024 |
| Benchmark | Score | Source | Published |
|---|---|---|---|
| LMArena EloLMArena | 1224 Elo · #183 of 191 | LMArena | Nov 19, 2026 |
Apps
Use Llama 3 8B Instruct in the tools you already run. Each guide points the app at the Respan gateway.
Setup guides for more apps are in the integration docs.
Llama 3 8B Instruct FAQ
Short answers, built from the data on this page.
Llama 3 8B Instruct is a model from Meta, released on Apr 18, 2024. Meta's Llama 3 model is rolled out in varying sizes, with the 8B instruct-tuned variant particularly optimized for high-quality dialogue purposes. The performance of Llama 3 has been found to be strong when compared to leading closed-source models, backed by human evaluations. The model family includes 8B and 70B sizes, with pretrained and instruct-tuned variants suitable for assistant-like chat. Llama 3 leverages auto-regressive language modeling with optimized transformer architecture. Training data is a mix of publicly available online data, with a knowledge cutoff in March 2023 for 8B and December 2023 for 70B models. The models use Grouped-Query Attention for improved inference scalability.
On the Respan gateway, Llama 3 8B Instruct costs $0.30 per 1M input tokens and $0.60 per 1M output tokens on every provider, after gateway discounts.
Llama 3 8B Instruct is available through 2 providers: Bedrock and Concentrate.
Llama 3 8B Instruct accepts up to 8K tokens of context on its largest provider route.
Point any OpenAI-compatible SDK at https://gateway.respan.ai/v1 with your Respan API key and set the model to "llama-3-8b-instruct". To set it up with a coding agent, copy the agent setup prompt at the top of this page.
Llama 3 8B Instruct has 5 published benchmark scores, including HumanEval 62.2%, GSM8K 79.6%, and MMLU 68.4%. Each score links to where it was published; the data is collected by Respan.
Certified where it counts.
Built to meet the security and privacy standards that enterprise and healthcare teams require.
ISO 27001
The internationally recognized standard for information security management.
SOC 2
Secure, compliant management of your data across all of our systems.
GDPR
Operated under GDPR, the world's strictest standard for data privacy.
HIPAA
HIPAA compliant, with a BAA available for healthcare teams.