Llama 3.1 70B Instruct pricing & benchmarks
Llama 3.1 70B Instruct is available from Meta at $0.400 per million input tokens and $0.400 per million output tokens ($0.400 blended at 3:1). It accepts up to 131,072 tokens of context. Meta's latest class of model (Llama 3.1) launched with a variety of sizes & flavors. This 70B instruct-tuned version is optimized for high quality dialogue usecases. It has demonstrated strong...
Pricing
Per million tokens · OpenRouter
Input$0.400
Output$0.400
Blended (3:1)$0.400
Specification
meta-llama/llama-3.1-70b-instruct
Context window131K
Max output16K
ReleasedJul 2024
LicenceOpen weights
Modalitiesin:text, out:text
Benchmarks
Independently run — not vendor self-reports. Every score links its source.
| Benchmark | Score | Config | Run | Source |
|---|---|---|---|---|
| GPQA Diamond | 44.2% ±2.5 | — | 2025-01-27 | Independently run · Epoch AI |
| OTIS Mock AIME | 3.6% ±1.7 | — | 2025-02-25 | Independently run · Epoch AI |
Scores published by Epoch AI, 'AI Benchmarking Hub', used under CC BY 4.0.
Cheaper models that score at least as well on GPQA Diamond
Strictly lower blended price AND an equal-or-higher GPQA Diamond score. One benchmark is one dimension — a model that wins here may still be weaker at your specific task.
| Model | Provider | Blended $/1M | GPQA Diamond | |
|---|---|---|---|---|
| Mistral Small 3 | Mistral AI | $0.058 (−86%) | 45.3% | Compare → |
| gpt-oss-120b | OpenAI | $0.070 (−82%) | 75.8% | Compare → |
| Phi 4 | Microsoft | $0.088 (−78%) | 56.1% | Compare → |
| DeepSeek V4 Flash 0731 | DeepSeek | $0.113 (−72%) | 91.0% | Compare → |
| GPT-5 Nano | OpenAI | $0.138 (−66%) | 69.4% | Compare → |
Where to run Llama 3.1 70B Instruct
3 hosts · 2.0× between cheapest and dearest. Blended at 3:1. ⚠ marks an option serving less context than the best available, or re-pricing above a prompt-length threshold. Bracketed labels are the provider's own service tier or region.
| Host | Blended $/1M | In / Out | Context | Weights | Uptime 24h |
|---|---|---|---|---|---|
| DeepInfra (turbo) | $0.400 | $0.400 / $0.400 | 131K | fp8 | 98.8% |
| Amazon Bedrock | $0.720 | $0.720 / $0.720 | 131K | — | 99.8% |
| CoreWeave | $0.800 | $0.800 / $0.800 | 128K | bf16 | 100.0% |
Host prices and uptime from OpenRouter, synced 2026-08-08. Lower-precision weights (fp4, fp8) can cost less and score worse than the same model at bf16 — the score above is not host-specific.
Compare Llama 3.1 70B Instruct
Llama 3.1 70B Instruct vs Qwen3.8 Max
Llama 3.1 70B Instruct vs DeepSeek V4 Flash 0731
Llama 3.1 70B Instruct vs Claude Opus 5
Llama 3.1 70B Instruct vs Gemini 3.6 Flash
Llama 3.1 70B Instruct vs Kimi K3
Llama 3.1 70B Instruct vs GPT-5.6 Luna
Llama 3.1 70B Instruct vs GPT-5.6 Terra
Llama 3.1 70B Instruct vs GPT-5.6 Sol
FAQ
How much does Llama 3.1 70B Instruct cost?
Llama 3.1 70B Instruct costs $0.400 per million input tokens and $0.400 per million output tokens on OpenRouter. At a 3:1 input:output mix that blends to $0.400 per million tokens.
What is Llama 3.1 70B Instruct's context window?
Llama 3.1 70B Instruct accepts up to 131,072 tokens of context, and can generate up to 16,384 output tokens.
How good is Llama 3.1 70B Instruct on benchmarks?
Llama 3.1 70B Instruct scores 44.2% on GPQA Diamond, independently run and published by Epoch AI. 2 benchmark results are listed on this page, each with its run date and source.
Is there a cheaper model as good as Llama 3.1 70B Instruct?
Yes — 5 models in our catalogue cost less per token than Llama 3.1 70B Instruct and score at least as high on the same benchmark. The cheapest is Mistral Small 3 at $0.058 per million tokens blended.