Llama 3 Replicate Base: pricing, context window and benchmarks
Llama 3 Replicate Base is a open-weight model from Replicate. It offers a 65,536-token context window, costs $3.15 per million input tokens and $16.13 per million output tokens, and averages 72.1 across the index's benchmark suite.
Pricing
Llama 3 Replicate Base costs $3.15 per million input tokens and $16.13 per million output tokens. A blended 3:1 input-to-output mix works out to about $6.39 per million tokens. Cached input is listed at $0.36 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $3.15 | per 1M tokens |
| Output | $16.13 | per 1M tokens |
| Cached input | $0.36 | per 1M tokens |
| Blended 3:1 | $6.39 | estimate per 1M tokens |
| 1M in + 1M out | $19.28 | estimate |
Benchmarks
Scores below are modelled estimates on a 0–100 scale. Use them to compare relative position, not as official vendor results.
| Benchmark | Score | Rank | Percentile | Relative |
|---|---|---|---|---|
| MMLU-Pro | 72.9 | 368 | 39th | |
| GPQA Diamond | 66.4 | 339 | 44th | |
| SWE-bench Verified | 62.2 | 274 | 54th | |
| HumanEval | 74.0 | 358 | 40th | |
| MATH-500 | 75.7 | 310 | 48th | |
| AIME 2025 | 69.9 | 259 | 57th | |
| LiveCodeBench | 66.8 | 256 | 57th | |
| IFEval | 81.3 | 295 | 51th | |
| TruthfulQA | 76.0 | 306 | 49th | |
| HellaSwag | 82.0 | 347 | 42th | |
| MMMU | 66.0 | 373 | 38th | |
| Arena Elo | 1376.7 | 285 | 53th |
What Llama 3 Replicate Base is good for
Llama 3 Replicate Base is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $3.15 per million input tokens (hosted list price) and a 65,536-token context window, it is a practical option for teams that want more control than a closed API allows.
Strengths
- Input price of $3.15 per million tokens is at or below the index median.
- Open weights make self-hosting, fine-tuning and data-residency review possible.
- Accepts more than text, including image input where the provider supports it.
Trade-offs
- Average benchmark score is below the index median; test it on your own task.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Llama 3.1 Base | Meta AI | 32768 | $3.39 | 72.2 |
| Seed-OSS Base | ByteDance | 16384 | $3.14 | 72.2 |
| Llama 3 Replicate Base | Replicate | 65536 | $3.15 | 72.1 |
| Claude Sonnet 4.5 Nano | Anthropic | 8192 | $8.93 | 72.0 |
| Mixtral Groq Chat | Groq | 8192 | $3.09 | 72.0 |
| o4 Nano | OpenAI | 65536 | $8.61 | 72.0 |
FAQ
How much does Llama 3 Replicate Base cost?
Llama 3 Replicate Base is listed at $3.15 per million input tokens and $16.13 per million output tokens. Cached input is $0.36 per million tokens. A blended 3:1 input-to-output mix is about $6.39 per million tokens.
What is the context window of Llama 3 Replicate Base?
Llama 3 Replicate Base has a context window of 65,536 tokens and can emit up to 65,536 output tokens per request in this index.
Is Llama 3 Replicate Base open weights?
Yes. Llama 3 Replicate Base is listed as an open-weight model under a open weights licence.
How does Llama 3 Replicate Base score on benchmarks?
Llama 3 Replicate Base averages 72.1 across the scored benchmarks in this index, ranking #326 of 600 models. Its strongest relative result is Arena Elo at 1376.7.
What is Llama 3 Replicate Base best for?
Llama 3 Replicate Base is an open-weight model, so it can be self-hosted or run through a third-party inference provider. That matters for data residency, fine-tuning and long-run cost control. At $3.15 per million input tokens (hosted list price) and a 65,536-token context window, it is a practical option for teams that want more control than a closed API allows.