Head-to-Head Pricing Benchmark
Side-by-side pricing and context window comparison for production model selection.
Default Recommendation (120M input + 60M output)
Qwen: Qwen3 235B A22B Instruct 2507 is lower-cost for the default monthly workload scenario.
Adjust the workload in the calculator below to see a live recommendation for your usage.
| Metric | Baidu: ERNIE 4.5 21B A3B | Qwen: Qwen3 235B A22B Instruct 2507 |
|---|---|---|
| Developer | baidu | qwen |
| Context Window | 120,000 | 262,144 |
| Max Output Tokens | 8,000 | 16,384 |
| Input Cost / 1M Tokens | $0.0700 | $0.0710 |
| Output Cost / 1M Tokens | $0.2800 | $0.1000 |
| Projected Monthly Cost | $25 | $15 |
| Vision | ❌ No | ❌ No |
| Tool Calling | ✅ Yes | ✅ Yes |
| Structured Output | ❌ No | ✅ Yes |
| Reasoning | ❌ No | ❌ No |
| Web Search | ❌ No | ❌ No |
| Tokenizer | Other | Qwen3 |
| MMLU Score | N/A | 82.8 |
| GPQA | N/A | 75.3 |
Price History
Current Input / 1M
$0.0700
Current Output / 1M
$0.2800
Price History
Current Input / 1M
$0.0710
Current Output / 1M
$0.1000
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
100.0%
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
98.9%
Adjust your workload to see projected monthly costs.
Baidu: ERNIE 4.5 21B A3B
$25
per month
Qwen: Qwen3 235B A22B Instruct 2507
$15
per month
Lower costLive Recommendation
Qwen: Qwen3 235B A22B Instruct 2507 is lower-cost at 120M input + 60M output tokens/month.
Continue evaluation with more “A vs B pricing” decision pages.