Kolibri-1
Aleph Alpha's open-weight English and German reasoning model, a mixture of experts with 78.1B parameters (3.46B active). Apache 2.0 license.
- Params
- 78.1B, 3.46B active
- Context
- 1M tokens
- Cutoff
- Released
- License
- Apache 2.0
Hosted API pricing
How we price
Order
Providers are listed cheapest input rate first, then by output rate, a tie going to the creator's own listing. A provider that lists the model without a published rate sits at the end.
Prices
Base-tier, on-demand rates in USD per 1M tokens, converted at ECB reference rates where a provider publishes in another currency. Cached-input and batch rates show only where a provider publishes them. "Cheapest" marks the one provider with the lowest input and output pair; a tie wears no badge.
Cost
Input rate times the input tokens plus output rate times the output tokens, at the monthly volume set above the table. Cached-input, batch, long-context and reasoning-token billing are not modelled.
Transparency and funding
- Affiliates: Affiliate links are marked. We may earn a commission if you click them, but commissions never affect the order.
Every provider serving Kolibri-1, cheapest input first. Set your monthly volume to see what each would bill.
| Provider | Input /1M | Output /1M | Cost at 10M in + 2M out | Link |
|---|---|---|---|---|
|
Aleph Alpha
Creator
|
Open weights, no hosted price | Visit website | ||
Heads up: Base-tier, on-demand rates per 1M tokens; cached, batch and long-context tiers excluded. A provider may serve a shorter context or a quantized build than the creator's release. Verify before provisioning. More on how we price.
Estimated cost to self-host
Kolibri-1 needs about 45 GB of GPU memory at 4-bit with 32K context. The cheapest rental that fits is 4× RTX 4060 Ti at about $403 a month.
| Precision | Memory | Cheapest fit (32K context) | Cost /mo |
|---|---|---|---|
| 4-bit INT4 / FP4 | 45 GB |
4× RTX 4060 Ti
|
$403 |
| 8-bit FP8 / INT8 | 81 GB |
8× RTX 3060
|
$461 |
| 16-bit FP16 / BF16 | 159 GB |
8× RTX 3090
|
$1,152 |
Estimates based on median on-demand rates for Nvidia GPUs. Memory is weights plus KV cache for one request, using FP8 KV cache where supported and FP16/BF16 otherwise. No guarantee of runtime support, usable performance, or that a matching quantized build exists. Pricing methodology.
Capabilities
What Kolibri-1 accepts and can do, as published by Aleph Alpha.
- Text Accepts and generates natural-language text.
- Reasoning Thinks before it answers, always on or as a switchable mode.
- Function calling Connect to external tools, APIs, and systems.
Common questions
What is Kolibri-1 good for?
German and English document work, question answering and tool-calling agents where data has to stay on infrastructure you control. Aleph Alpha built it for public administration, industry and aerospace, and its Apache 2.0 weights fit a single high-memory GPU at 8-bit.
When is Kolibri-1 not a good fit?
Image, audio or video input: it is text only, and no maximum output is published. Languages other than English and German, which Aleph Alpha does not target.
What is the cheapest way to run Kolibri-1?
No provider we track hosts it; self-hosting on the cheapest rental that fits (4x RTX 4060 Ti, $0.56 per hour) costs about $403 a month.
Can I self-host Kolibri-1?
Yes. The weights are Apache 2.0 licensed. At 4-bit it needs about 45 GB of GPU memory, which starts at roughly $403 a month on the cheapest rental that fits.