Head-to-head showdown: AI21 Jamba 1.5 Large ($2.00 in / $8.00 out per 1M) vs OpenRouter Auto-Best Router ($1.00 in / $3.00 out per 1M). OpenRouter Auto-Best Router is 2.5× cheaper across standard token mixes, with 256K vs 1M context windows.
jamba-1.5-large · ai21
openrouter/auto · openrouter
Benchmark
| Workload Scenario | AI21 Jamba 1.5 Large | OpenRouter Auto-Best Router | Price Delta |
|---|---|---|---|
| 1M input tokens (raw text) | $2.00 | $1.00 | +$1.00 |
| 1M output tokens (generation) | $8.00 | $3.00 | +$5.00 |
| 1M tokens · 70% input / 30% output mix | $3.80 | $1.60 | +$2.20 |
| Standard chat turn (4K in / 800 out, 50% cached) | $0.0144 | $0.0064 | +$0.008 |
| Monthly scale (10K requests / day) | $4,320.00 | $1,920.00 | +$2,400.00 |
Negative difference = AI21 Jamba 1.5 Large is cheaper. Positive = OpenRouter Auto-Best Router is cheaper.
Scaling Curve
Total cost of a token volume at 70/30 input/output split (uncached). Log-log scale.
Verdict
OpenRouter Auto-Best Router is cheaper on both input and output rates, so it costs less at every input/output mix. Price alone still isn't the whole decision: capability, latency and context limits (AI21 Jamba 1.5 Large: 256K, OpenRouter Auto-Best Router: 1M) may justify the premium for your task.
FAQ
On input, OpenRouter Auto-Best Router is cheaper ($1.00/M vs $2.00/M). On output, OpenRouter Auto-Best Router is cheaper ($3.00/M vs $8.00/M). The same model is cheaper on both sides, so it wins at every mix.
A chat-style request (4,000 input + 800 output tokens, 50% cached) costs $0.0144 on AI21 Jamba 1.5 Large and $0.0064 on OpenRouter Auto-Best Router — OpenRouter Auto-Best Router is 2.3× more expensive for that workload.
AI21 Jamba 1.5 Large supports 256,000 tokens (4.1K max output); OpenRouter Auto-Best Router supports 1,000,000 (32.8K max output). OpenRouter Auto-Best Router fits 3.9× more context, which matters for long documents and agents.