ERNIE 4.0 Max: pricing, context window and benchmarks
ERNIE 4.0 Max is a multimodal model from Baidu. It offers a 32,768-token context window, costs $12.40 per million input tokens and $53.07 per million output tokens, and averages 78.9 across the index's benchmark suite.
Pricing
ERNIE 4.0 Max costs $12.40 per million input tokens and $53.07 per million output tokens. A blended 3:1 input-to-output mix works out to about $22.57 per million tokens. Cached input is listed at $1.69 per million tokens.
| Item | Price | Notes |
|---|---|---|
| Input | $12.40 | per 1M tokens |
| Output | $53.07 | per 1M tokens |
| Cached input | $1.69 | per 1M tokens |
| Blended 3:1 | $22.57 | estimate per 1M tokens |
| 1M in + 1M out | $65.47 | 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 | 80.8 | 152 | 75th | |
| GPQA Diamond | 76.4 | 110 | 82th | |
| SWE-bench Verified | 68.0 | 145 | 76th | |
| HumanEval | 83.5 | 142 | 76th | |
| MATH-500 | 84.6 | 115 | 81th | |
| AIME 2025 | 76.5 | 135 | 78th | |
| LiveCodeBench | 70.1 | 188 | 69th | |
| IFEval | 85.9 | 161 | 73th | |
| TruthfulQA | 81.6 | 138 | 77th | |
| HellaSwag | 88.7 | 120 | 80th | |
| MMMU | 72.0 | 242 | 60th | |
| Arena Elo | 1405.1 | 194 | 68th |
What ERNIE 4.0 Max is good for
ERNIE 4.0 Max is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $12.40 per million input tokens, it sits in the mid-range of the index, and the 32,768-token context window is enough for large files and multi-file diffs.
Strengths
- Accepts more than text, including image input where the provider supports it.
- Coding benchmarks are a relative strength within its tier.
Trade-offs
- Proprietary licence: you cannot download the weights or self-host the model.
- Premium input pricing at $12.40 per million tokens.
How it compares
| Model | Provider | Context | Input / 1M | Avg score |
|---|---|---|---|---|
| Sonar Pro | Perplexity | 128000 | $13.01 | 79.1 |
| Claude Opus 4.5 Mini | Anthropic | 65536 | $12.72 | 79.0 |
| QwQ Instruct | Alibaba Cloud | 65536 | $4.62 | 79.0 |
| Sonar Pro Turbo | Perplexity | 128000 | $13.55 | 79.0 |
| ERNIE 4.0 Max | Baidu | 32768 | $12.40 | 78.9 |
| GLM-4.5 Instruct | Zhipu AI | 200000 | $4.70 | 78.8 |
FAQ
How much does ERNIE 4.0 Max cost?
ERNIE 4.0 Max is listed at $12.40 per million input tokens and $53.07 per million output tokens. Cached input is $1.69 per million tokens. A blended 3:1 input-to-output mix is about $22.57 per million tokens.
What is the context window of ERNIE 4.0 Max?
ERNIE 4.0 Max has a context window of 32,768 tokens and can emit up to 8,192 output tokens per request in this index.
Is ERNIE 4.0 Max open weights?
No. ERNIE 4.0 Max is a proprietary model with a proprietary licence, so you cannot download the weights.
How does ERNIE 4.0 Max score on benchmarks?
ERNIE 4.0 Max averages 78.9 across the scored benchmarks in this index, ranking #152 of 600 models. Its strongest relative result is Arena Elo at 1405.1.
What is ERNIE 4.0 Max best for?
ERNIE 4.0 Max is tuned toward coding work. It is a reasonable candidate for code generation, review, debugging and repository-scale tasks, especially when paired with a retrieval layer. At $12.40 per million input tokens, it sits in the mid-range of the index, and the 32,768-token context window is enough for large files and multi-file diffs.