Catalog prices for Metawareness models — then a comparison with other Swiss-hosted providers. We name the models that appear on the overview intelligence chart, not the competing platforms.
Rates match the live model registry. Swiss-hosted rows run on our hardware or a Swiss provider; routed rows use the Metawareness synonym over cloud upstreams.
| Model | Hosting | Intelligence index | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|---|
| Nymara Pro 1.3 | Routed | 45 | $6 | $18 |
| Nymara Flash 1.3 | Routed | 42 | $4 | $12 |
| Duskara 1.1 | Routed | 40 | $6 | $18 |
| Luneth Next 1.8 | Swiss-hosted | 40 | $2 | $6 |
| Myrion 1.6 | Routed | 38 | $4 | $12 |
| Luneth 1.8 | Swiss-hosted | 34 | $2 | $6 |
| Qwen3.8 27B | Swiss-hosted | 34 | $2 | $6 |
| Kimi 2.6 | Swiss-hosted | 27 | $0.70 | $3.50 |
Prices are USD per 1 000 000 tokens. CHF catalog entries are shown converted at ≈ 1.15 USD/CHF for a single currency view.
Other Swiss-hosted providers bill differently: some meter tokens at about $2.50 input / $4 output per 1M; others sell monthly plans from CHF 1 000+ per model (CHF 2 500–2 900 for larger / all-models tiers) for roughly 100M-token-scale allowances.
| Provider | Hosting | Input / 1M | Output / 1M | How billed |
|---|---|---|---|---|
| Metawareness — Luneth / local Swiss models | Swiss-hosted | $2 | $6 | Pay only for tokens used |
| Other Swiss-hosted provider (usage API) | Swiss-hosted | $2.50 | $4 | Usage-priced Swiss API |
| Other Swiss-hosted provider (monthly plans) | Swiss-hosted | ≈ $11.50 blended | ≈ $11.50 blended | Flat monthly plans from CHF 1 000 (often ≈ 100M tokens included); all-models / flagship tiers CHF 2 500–2 900/mo |
We meter tokens. Monthly Swiss seats are fine if you burn the whole allotment every month — expensive if you do not.
Our Swiss local stack is $2 / 1M input vs a typical Swiss usage API at $2.50 / 1M. Output is higher on our side ($6 vs $4) — the trade is intelligence, not the cheapest completion token.
CHF 1 000 / month for ~100M tokens is about $11.50 / 1M blended — before you split input and output. Higher monthly tiers push that number up further.
Other Swiss-hosted bars — Kimi K2.6, Qwen3.5 122B, Apertus 70B — sit below every Metawareness Swiss model. Expensive non-Swiss cloud models above or near our band (Opus, Fable, Astra, Sonnet, Kimi K3, GPT-5.6 Sol) cost far more per token for a residency posture you may not need. Prior-gen Opus 4.5 / 4.8 and Sonnet / GPT rows still bill like frontier models while landing closer to our intelligence band — the worst of both: pay Opus money, get mid-chart results. Some Swiss APIs are cheaper and weaker; some cloud APIs are smarter and much more expensive. Price the job, not the logo.
| Model | Hosting | Intelligence index | Input / 1M | Output / 1M | Notes |
|---|---|---|---|---|---|
| Claude Opus 5.5 | Other cloud | 58 | $4 | $20 | higher |
| Claude Sonnet 5.5 | Other cloud | 56 | $2 | $10 | higher |
| Claude Fable 5.1 | Other cloud | 53 | $10 | $50 | higher |
| GPT-6 Astra | Other cloud | 53 | $10 | $50 | higher |
| Claude Opus 5 | Other cloud | 52 | $5 | $25 | prior vs 5.5 · still frontier-priced |
| Claude Sonnet 5 | Other cloud | 50 | $2 | $10 | prior vs Sonnet 5.5 · similar band |
| GPT-5.6 Sol | Other cloud | 48 | $5 | $30 | similar / slightly higher |
| Claude Opus 4.8 | Other cloud | 46 | $5 | $25 | similar / slightly higher |
| Nymara Pro 1.3 (Metawareness) | Metawareness | 45 | $6 | $18 | Metawareness |
| GPT-5.6 Terra | Other cloud | 44 | $2.50 | $15 | mid GPT-5.6 tier · similar band |
| Kimi K3 | Other cloud | 44 | $3 | $15 | similar to Nymara Pro |
| Claude Opus 4.5 | Other cloud | 42 | $5 | $25 | similar to Nymara Flash |
| Nymara Flash 1.3 (Metawareness) | Metawareness | 42 | $4 | $12 | Metawareness |
| Duskara 1.1 (Metawareness) | Metawareness | 40 | $6 | $18 | Metawareness |
| GPT-5 | Other cloud | 40 | $1.25 | $10 | prior GPT line · similar / lower |
| Luneth Next 1.8 (Metawareness) | Swiss-hosted | 40 | $2 | $6 | Metawareness |
| Myrion 1.6 (Metawareness) | Metawareness | 38 | $4 | $12 | Metawareness |
| Luneth 1.8 (Metawareness) | Swiss-hosted | 34 | $2 | $6 | Metawareness |
| Kimi K2.6 | Other Swiss-hosted | 27 | $0.70 | $3.50 | other Swiss-hosted |
| Qwen3.5 122B | Other Swiss-hosted | 16 | $0.46 | $3.68 | other Swiss-hosted |
| Apertus 70B | Other Swiss-hosted | 5 | $0.805 | $2.875 | other Swiss-hosted |
Apertus 70B and Qwen3.5 122B can undercut on raw $/1M. On the chart they sit near the bottom of Swiss-hosted options. Cheap is only a win if quality holds.
Claude Opus 4.8 and 4.5 at $5 / $25 (same prior Opus list price) and Sonnet 5 at $2 / $10 sit closer to our band than Opus 5.5 — but still cost a multiple of Luneth ($2 / $6) without Swiss-hosted residency.
Claude Fable 5.1 and GPT-6 Astra run about $10 / $50 per 1M — several times Nymara or Luneth — for a non-Swiss path. Worth it only when the extra intelligence index actually moves the decision.
Metawareness models occupy the top of the Swiss column on the intelligence chart at catalog rates that undercut frontier and prior-gen cloud pricing. Swiss residency without the Opus tax.
Competitor figures are representative public list prices (Swiss usage / monthly plans and major cloud APIs). Intelligence ranks refer to the chart on the overview. Not a contractual offer — verify current pricing before you buy.