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 | Qwen: Qwen3 235B A22B Instruct 2507 | Baidu: ERNIE 4.5 21B A3B |
|---|---|---|
| Developer | qwen | baidu |
| Context Window | 262,144 | 120,000 |
| Max Output Tokens | — | 8,000 |
| Input Cost / 1M Tokens | $0.0710 | $0.0700 |
| Output Cost / 1M Tokens | $0.1000 | $0.2800 |
| Projected Monthly Cost | $15 | $25 |
| Vision | ❌ No | ❌ No |
| Tool Calling | ✅ Yes | ✅ Yes |
| Structured Output | ✅ Yes | ❌ No |
| Reasoning | ❌ No | ❌ No |
| Web Search | ❌ No | ❌ No |
| Tokenizer | Qwen3 | Other |
| MMLU Score | 82.8 | N/A |
| GPQA | 75.3 | N/A |
Price History
Current Input / 1M
$0.0710
Current Output / 1M
$0.1000
Price History
Current Input / 1M
$0.0700
Current Output / 1M
$0.2800
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
96.3%
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
100.0%
Adjust your workload to see projected monthly costs.
Qwen: Qwen3 235B A22B Instruct 2507
$15
per month
Lower costBaidu: ERNIE 4.5 21B A3B
$25
per month
Live 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.