Head-to-Head Pricing Benchmark
Side-by-side pricing and context window comparison for production model selection.
Default Recommendation (120M input + 60M output)
Meta: Llama 3 8B 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 | IBM: Granite 4.0 Micro | Meta: Llama 3 8B Instruct |
|---|---|---|
| Developer | ibm-granite | meta-llama |
| Context Window | 131,000 | 8,192 |
| Max Output Tokens | — | 16,384 |
| Input Cost / 1M Tokens | $0.0170 | $0.0300 |
| Output Cost / 1M Tokens | $0.1100 | $0.0400 |
| Projected Monthly Cost | $8.64 | $6.00 |
| Vision | ❌ No | ❌ No |
| Tool Calling | ❌ No | ✅ Yes |
| Structured Output | ❌ No | ✅ Yes |
| Reasoning | ❌ No | ❌ No |
| Web Search | ❌ No | ❌ No |
| Tokenizer | Other | Llama3 |
| MMLU Score | N/A | 47.6 |
| GPQA | N/A | 25.9 |
Price History
Current Input / 1M
$0.0170
Current Output / 1M
$0.1100
Price History
Current Input / 1M
$0.0300
Current Output / 1M
$0.0400
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
100.0%
Performance History
Current TPS
0.00
Current Latency
0ms
Uptime
100.0%
Adjust your workload to see projected monthly costs.
IBM: Granite 4.0 Micro
$8.64
per month
Meta: Llama 3 8B Instruct
$6.00
per month
Lower costLive Recommendation
Meta: Llama 3 8B Instruct is lower-cost at 120M input + 60M output tokens/month.
Continue evaluation with more “A vs B pricing” decision pages.