IBM Granite 4.2 8B vs Qwen3 14B
Side by side from live data: prices, limits, who serves each model and how reliably, and how they do on our chess arena.
In short
- Granite 4.2 8B is 1.4× cheaper at a typical mix of 3 input tokens per output token.
- Granite 4.2 8B accepts longer prompts: 131K vs 41K tokens.
- Granite 4.2 8B output costs $0.15 per 1M via CoreWeave, against the list price of $0.25.
- Granite 4.2 8B is the newer model (released 2026-08-31).
| Granite 4.2 8B | Qwen3 14B | |
|---|---|---|
| Vendor | IBM | Qwen |
| Released | 2026-08-31 | 2025-04-28 |
| Input / 1M | $0.06 | $0.12 |
| Output / 1M | $0.25 | $0.24 |
| Cached input / 1M | $0.015 | — |
| Blended / 1M (3:1) | $0.107 | $0.15 |
| Context | 131K | 41K |
| Max output | 118K | 16K |
| Accepts | text | text |
| Reasoning | Yes | Yes |
| Providers on OpenRouter | 2 | 2 |
| Cheapest provider, output | $0.15 | $0.22 |
| Best provider uptime, 24 h | 100% | 100% |
| Chess arena rating | not playing | not playing |
| Vendor API status | — | Operational |
Cost calculator
Granite 4.2 8B
$7.35 per month$0.24 per day
Qwen3 14B
$10.80 per month$0.36 per day
Granite 4.2 8B saves $3.45 a month on this workload.
At list prices per 1M tokens; prompt caching and batch discounts are not included.
Compare Granite 4.2 8B with