Meta model product
Llama 3.1 70B
1 70B Instruct is a large language model optimized for multilingual dialogue use cases.
Updated Aug 12, 2026. Default version: Llama 3.1 70B Instruct
LLMBoard score13.2Llama 3.1 70B Instruct
Coverage100%16 benchmark families
Context window128KTokens
Official input priceN/AOfficial price unavailable
Llama 3.1 70B Specifications
Technical details for the model's default version.
- Version
- Llama 3.1 70B Instruct
- Released
- Jul 23, 2024
- Knowledge cutoff
- Unknown
- Parameters
- 70B
- Context window
- 128K
- Max output
- 128K
- Inputs
- text
- Outputs
- text
- Open weights
- No
- License
- Llama 3.1 Community License
Llama 3.1 70B Capability Profile
This profile uses the model's current scored version. Arena ratings and prices are shown separately.
Llama 3.1 70B Benchmark Results
Benchmark scores for Llama 3.1 70B Instruct.
| GSM-8K (CoT) | 95.1% | 01 | 2 | 100.0% | C | |
| MATH (CoT) | 68.0% | 01 | 6 | 100.0% | C | |
| MBPP ++ base version | 86.0% | 01 | 1 | 100.0% | C | |
| API-Bank | 90.0% | 02 | 3 | 50.0% | C | |
| BFCL | 84.8% | 02 | 11 | 90.0% | C | |
| Gorilla Benchmark API Bench | 29.7% | 02 | 3 | 50.0% | C | |
| MMLU (CoT) | 86.0% | 02 | 3 | 50.0% | C | |
| Multilingual MGSM (CoT) | 86.9% | 02 | 3 | 50.0% | C | |
| Multipl-E HumanEval | 65.5% | 02 | 3 | 50.0% | C | |
| Multipl-E MBPP | 62.0% | 02 | 3 | 50.0% | C | |
| Nexus | 56.7% | 02 | 4 | 66.7% | C | |
| ARC-C | 94.8% | 04 | 34 | 90.9% | C | |
| DROP | 79.6% | 14 | 30 | 55.2% | C | |
| IFEval | 87.5% | 32 | 65 | 51.6% | C | |
| HumanEval | 80.5% | 45 | 66 | 32.3% | C | |
| MMLU | 83.6% | 47 | 100 | 53.5% | C | |
| MMLU-Pro | 66.4% | 98 | 129 | 24.2% | C | |
| GPQA | 41.7% | 201 | 234 | 14.2% | C |
Llama 3.1 70B Arena Results
Preference and agent-evaluation results for the default version.
| text | overall | 260 | 1261.0 | 55,240 | N/A | |
| text style control | overall | 261 | 1293.3 | 55,240 | N/A |
Llama 3.1 70B Runtime Performance
Provider-specific output speed and catalog latency for Llama 3.1 70B Instruct. Runtime does not affect the capability score.
| Cerebras | 1,204 tok/s | 0.2 s | 128K | 128K | |
| Groq | 250 tok/s | 0.5 s | 128K | 128K | |
| Hyperbolic | 100 tok/s | 0.5 s | 128K | 128K | |
| Bedrock | 100 tok/s | 0.5 s | 128K | 128K | |
| Together | 94 tok/s | 0.5 s | 128K | 128K | |
| Sambanova | 74 tok/s | 0.5 s | 128K | 128K | |
| Lambda | 42 tok/s | 0.5 s | 128K | 128K | |
| Fireworks | 32 tok/s | 0.5 s | 128K | 128K | |
| DeepInfra | 25 tok/s | 0.5 s | 128K | 128K |
Output Speed is generated output tokens received per second. Catalog latency is reported separately from observed provider TTFT.
Llama 3.1 70B Pricing
Official vendor API pricing appears first, followed by individual provider offers.
- Official API
- N/A
- Official provider
- N/A
- Lowest third-party
- From $0.40 input, $0.40 output per 1M via OpenRouter
- Tracked offerings
- 5
| Nvidia | meta/llama-3.1-70b-instruct | global | N/A | N/A | 128K | |
| OpenRouter | meta-llama/llama-3.1-70b-instruct | global | $0.40 | $0.40 | 131.1K | |
| LLM Gateway | llama-3.1-70b-instruct | global | $0.72 | $0.72 | 128K | |
| Amazon Bedrock | meta.llama3-1-70b-instruct-v1:0 | global | $0.72 | $0.72 | 128K | |
| Vercel AI Gateway | meta/llama-3.1-70b | global | $0.72 | $0.72 | 128K |
Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.
Llama 3.1 70B Versions
Available versions of this model. The score column identifies the version used in the overall ranking.
| Llama 3.1 70B Instruct | 13.2 | 70B | 128K | 128K | No | Llama 3.1 Community License |
What is Llama 3.1 70B?
Key information about Llama 3.1 70B and its available data.
1 70B Instruct is a large language model optimized for multilingual dialogue use cases. It outperforms many available open source and closed chat models on common industry benchmarks.
Data as of 2026-08-11.
Llama 3.1 70B vs nearby models
Open a comparison with the three ranked models immediately above and below this model.
Models similar to Llama 3.1 70B
Recommendations prioritize the same model type and family, then the closest LLMBoard score.
ME
Llama 4 Scout
Meta
ME
Llama 3.2 90B
Meta
ME
Llama 3.3 70B
Meta
ME
Llama 3.1 405B
Meta
ME
Llama 4 Maverick
Meta
ME
Llama 3.2 3B
Meta
FAQ
Common questions about Llama 3.1 70B.
When was Llama 3.1 70B released?
Llama 3.1 70B's default version was released on Jul 23, 2024.
How much does Llama 3.1 70B cost?
No official standard PAYG price is currently available for Llama 3.1 70B. The lowest tracked third-party offer starts at $0.40 input and $0.40 output via OpenRouter.
Who created Llama 3.1 70B?
Llama 3.1 70B was created by Meta.
What is the context window for Llama 3.1 70B?
The default version has a 128K token context window.
Is Llama 3.1 70B open weight?
No. The default version is not marked as having publicly available weights.
How many API providers offer Llama 3.1 70B?
5 provider offerings are linked to the default version.
What models should I compare Llama 3.1 70B with?
Nearby ranked alternatives include Llama 4 Scout, Phi 4 Mini Reasoning, Nova Lite.