Head-to-Head Pricing Benchmark
Side-by-side pricing and context window comparison for production model selection.
Default Recommendation (120M input + 60M output)
Qwen: Qwen2.5 VL 72B Instruct 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: Qwen2.5 VL 72B Instruct | Anthropic: Claude 3.5 Haiku |
|---|---|---|
| Developer | qwen | anthropic |
| Context Window | 32,000 | 200,000 |
| Max Output Tokens | — | 8,192 |
| Input Cost / 1M Tokens | $0.2500 | $0.8000 |
| Output Cost / 1M Tokens | $0.7500 | $4.00 |
| Projected Monthly Cost | $75 | $336 |
| Vision | ✅ Yes | ✅ Yes |
| Tool Calling | ❌ No | ✅ Yes |
| Structured Output | ✅ Yes | ❌ No |
| Reasoning | ❌ No | ❌ No |
| Web Search | ❌ No | ❌ No |
| Tokenizer | Qwen | Claude |
| MMLU Score | 72.0 | N/A |
| GPQA | 49.1 | N/A |
Price History
Current Input / 1M
$0.2500
Current Output / 1M
$0.7500
Price History
Current Input / 1M
$0.8000
Current Output / 1M
$4.00
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
57.5%
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
99.9%
Adjust your workload to see projected monthly costs.
Qwen: Qwen2.5 VL 72B Instruct
$75
per month
Lower costAnthropic: Claude 3.5 Haiku
$336
per month
Live Recommendation
Qwen: Qwen2.5 VL 72B Instruct is lower-cost at 120M input + 60M output tokens/month.
Continue evaluation with more “A vs B pricing” decision pages.