Comparison LLM base token rates and Jev price checked 17 October 2026 against official sources. Other tiers and long-context rates can differ. Change the calculator rates to match your account.
Jev compared with the models you would otherwise use
Jev bills 0.042 dollars per million input tokens and nothing for output. An LLM bills both. These pages work the gap out per model, and each one says plainly where the LLM is still the better choice.
One workload, every model
Support triage: 100K tickets a day, a 500-token ticket, 4 questions answered in one call, and 50 output tokens per ticket for the LLM. 30 days, list prices.
| Model | Per month | vs Jev |
|---|---|---|
| Jev 1.13 | $126.00 | 1x |
| GPT-5.6 Luna | $624.00 | 5.0x |
| Claude Haiku 4.5 | $2,970 | 24x |
| Claude Sonnet 5 | $5,940 | 47x |
| GPT-5.6 Terra | $6,240 | 50x |
| GPT-5.6 Sol | $11,880 | 94x |
| Claude Opus 5 | $14,850 | 118x |
| Claude Fable 5.1 | $29,700 | 236x |
| GPT-6 Astra | $29,700 | 236x |
Jev includes the ~260 tokens each request is measured to bill; LLM rows count your content only, so they are a floor. Change the workload.
Per-model pages
Luna is the cheapest frontier-family LLM most teams reach for when classifying. Here is what the same workload costs on each, with output tokens counted.
GPT-5.6 Luna: $0.2/M in · $1.2/M out | Jev 1.13: $0.042/M in · free out
Haiku is the usual baseline for high-volume classification and routing. LiteLLM measured Jev classifying 5.43x faster at 96% lower cost on their router corpus.
Claude Haiku 4.5: $1/M in · $5/M out | Jev 1.13: $0.042/M in · free out
Sonnet is the mid-tier default in a lot of pipelines. On pure classification volume the gap is mostly output tokens, which Jev does not bill at all.
Claude Sonnet 5: $2/M in · $10/M out | Jev 1.13: $0.042/M in · free out
Terra is the step up teams reach for when Luna starts getting classifications wrong. It bills 10.0 times Luna's output rate, which is the half of the bill Jev does not have at all.
GPT-5.6 Terra: $2/M in · $12/M out | Jev 1.13: $0.042/M in · free out
Sol is usually chosen for accuracy rather than price, which makes it the clearest case for splitting a pipeline: Jev for the volume, Sol for the fraction that genuinely needs it.
GPT-5.6 Sol: $4/M in · $20/M out | Jev 1.13: $0.042/M in · free out
Opus is a poor fit for high-volume classification on price alone, so the interesting question is not which is cheaper but how much of your Opus spend is actually doing Opus-grade work.
Claude Opus 5: $5/M in · $25/M out | Jev 1.13: $0.042/M in · free out
Fable sits at the top of this comparison set. The gap against Jev is the largest here, which makes it the honest upper bound rather than the number to quote.
Claude Fable 5.1: $10/M in · $50/M out | Jev 1.13: $0.042/M in · free out
The frontier end of the comparison. This is where TypeSafe's headline savings ratios come from, so treat them as an upper bound rather than a forecast.
GPT-6 Astra: $10/M in · $50/M out | Jev 1.13: $0.042/M in · free out
Self-hosted alternative
The open-weight Laya model on a cloud GPU or your own Mac: the daily volume where it beats Jev’s bill, its latency measured on an M4 Pro, and the accuracy and context gaps to check first.
Get an email when Jev’s price or limits change
We check TypeSafe’s published price and limits every day, and will tell you when either moves.