Every model, one endpoint, one bill.
One key, one OpenAI-compatible URL, and 26 models behind it. Change the model string and nothing else about your code changes — including who you pay.
Or don’t pick one.
Nex 1 is Lobstack’s router: it scores each request and serves it from the catalogue below — the cheapest tier that clears the score, never above the ceiling your plan reaches. The response names the model that actually answered, so a routed call is auditable rather than a claim.
It has no rate of its own. It costs whatever it routed to, which is on the receipt. A model of our own is the phase after this one →
curl https://www.lobstack.ai/api/gateway/v1/chat/completions \
-H "Authorization: Bearer $LOBSTACK_API_KEY" \
-d '{ "model": "auto", "messages": [...] }'What the Gateway serves
Rates are what you pay, per million tokens, for traffic on our keys. List prices were last checked against each provider’s own page on 2026-09-08.
| Model | Model string | Provider | Tier | Context | $/M in | $/M out |
|---|---|---|---|---|---|---|
| GPT-OSS 20B | gpt-oss-20b | Groq | Nano | 131K | $0.094 | $0.375 |
| Ministral 3 8B | ministral-8b | Mistral | Nano | 256K | $0.188 | $0.188 |
| GPT-5.6 Luna | gpt-5.6-luna | OpenAI | Nano | 1.05M | $0.250 | $1.50 |
| Gemini 3.1 Flash Lite | gemini-3.1-flash-lite | Nano | 1.05M | $0.313 | $1.88 | |
| GPT-OSS 120B | gpt-oss-120b | Groq | Small | 131K | $0.188 | $0.750 |
| Mistral Small 4 | mistral-small-4 | Mistral | Small | 256K | $0.188 | $0.750 |
| Qwen 3.8 Flash | qwen3.8-flash | Alibaba | Small | 1M | $0.188 | $0.587 |
| Claude Haiku 4.5 | claude-haiku-4-5 | Anthropic | Small | 200K | $1.25 | $6.25 |
| Grok Build 0.1 | grok-build-0.1 | xAI | Small | 256K | $1.25 | $2.50 |
| DeepSeek V4 Flash | deepseek-v4-flash | DeepSeek | Standard | 1M | $0.338* | $1.38 |
| Qwen 3.7 Plus | qwen3.7-plus | Alibaba | Standard | 1M | $0.500 | $2.00 |
| Gemini 3.8 Flash | gemini-3.8-flash | Standard | 1.05M | $0.938 | $4.69 | |
| Grok 4.3 | grok-4.3 | xAI | Standard | 1M | $1.56 | $3.13 |
| Claude Sonnet 5 | claude-sonnet-5 | Anthropic | Standard | 1M | $2.50 | $12.50 |
| Mistral Medium 3.5 | mistral-medium-3.5 | Mistral | Premium | 256K | $1.88 | $9.38 |
| Gemini 3.1 Pro | gemini-3.1-pro | Premium | 1.05M | $2.50 | $15.00 | |
| GPT-5.6 Terra | gpt-5.6-terra | OpenAI | Premium | 1.05M | $2.50 | $15.00 |
| Grok 4.6 | grok-4.6 | xAI | Premium | 500K | $2.50 | $7.50 |
| Grok 4.5 | grok-4.5 | xAI | Premium | 500K | $2.50 | $7.50 |
| Qwen 3.8 Max | qwen3.8-max | Alibaba | Premium | 1M | $2.50 | $7.50 |
| Kimi K3 | kimi-k3 | Moonshot | Premium | 1.05M | $3.75 | $18.75 |
| DeepSeek V4 Pro | deepseek-v4-pro | DeepSeek | Flagship | 1M | $1.38* | $5.50 |
| GPT-5.6 Sol | gpt-5.6 | OpenAI | Flagship | 1.05M | $5.00 | $25.00 |
| Claude Opus 5 | claude-opus-5 | Anthropic | Flagship | 1M | $6.25 | $31.25 |
| Claude Fable 5.1 | claude-fable-5-1 | Anthropic | Flagship | 1M | $12.50 | $62.50 |
| GPT-6 Astra | gpt-6-astra | OpenAI | Flagship | 1.05M | $12.50 | $62.50 |
* This provider publishes no list price we could confirm on 2026-09-08, so the rate is our best reading of it. It is marked rather than quietly presented as verified.
How a rate becomes a bill
Tokens in times the input rate, tokens out times the output rate. Every call returns the model that served it, the token counts and the cost — so the figure on your bill is one you can rebuild from the responses you already have.
That cost draws down the model spend included in your plan. When it runs out you top up; nothing stops mid-sentence and nothing is billed at a rate other than the one in the table above.
On BYOK these rates do not apply at all. Your provider bills your tokens at your rates, and we price none of them — you pay us for the routing, the receipts and the Console, metered in requests.