Kimi K2 0905
Moonshot AI · released Sep 3, 2025 · moonshotai/Kimi-K2-Instruct-0905
- Type
- Open weightsCustom licence
- Params
- 1T
- Context
- 262K
32B active per word · about 197K words of context · download allowed, licence restricts use
Our take
Written Sep 1, 2026Kimi K2 0905 is a trillion-parameter mixture-of-experts text model from Moonshot AI with 32 billion active per token. It handles up to 262,144 tokens in a single request and carries a custom restricted licence rather than a permissive one.
Choose this when you need very large model capacity with controlled inference cost, or for long-context text work at 262K tokens. It suits budget-conscious high-volume inference where the cheapest tracked offers sit well below the fastest host. Skip it if you need measured quality scores, multimodal input, or a licence you can freely redistribute or fine-tune.
The case for it
- Over one trillion total parameters with only 32 billion active per token — roughly 32× parameter multiplication without proportional compute cost.
- 262,144-token request limit, among the longest we track.
- 68 tokens per second measured throughput on one host.
- Cheapest tracked offers sit well below the fastest host's price point.
The case against it
- Custom restricted licence — not Apache 2.0 or MIT — with redistribution and commercial terms more constrained than standard open weights.
- No benchmark scores in our data: chat, reasoning, coding and academic benchmarks all unverified.
- Text-only; no image, video or audio input or output.
How good is it?
We hold no score for this model.
So there is no figure here for everyday use, coding, agent work or writing. That is a gap in our data, not a low score.
Can you run it yourself?
GeForce RTX 4090 · 24 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Apple M1 Pro (16-core GPU) · 32 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Radeon RX 7900 XT · 20 GB
Too large for this card. The weights do not fit even with part of them offloaded to system memory.
Memory use by level
Against a 24 GB card.
What is quantisation? →This model on every device we track71 devicesThe Q4 build most people download, on each device: what the weights come to, how much context the memory leaves, and whether it runs. Smallest device that runs it first.
Check against your own machine → · Where to rent it hosted →
Or rent it from someone else
Prices checked 2 hours ago — each listing carries its own date.
- per 1M tokens
- $0.60 in / $2.50 out
- Context served
- 262K
- Throughput
- Not measured
| Provider | In / out per 1M tokens | Context | Throughput | Trains on prompts | Logs prompts | Zero retention |
|---|---|---|---|---|---|---|
| OpenRouterOpenRouter's own listing | $0.60 / $2.50checked 2 hours ago | 262K | not measured | Unknown | Unknown | Unknown |
| Novita AIfp8Direct and through OpenRouter | $0.60 / $2.50checked 2 hours ago | 262K98K max reply through OpenRouter | 42 tok/sthrough OpenRouter | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterNo | DirectUnknownThrough OpenRouterConfirmed |
Novita AIAcross the 2 listings we hold: 1 says it does not train on prompts (only through OpenRouter), 0 say they do and 1 does not say. 1 appears in the zero-retention registry we check (only through OpenRouter); the rest are unknown to us.
What each host's API supports
From the parameter list each endpoint publishes. Streaming is omitted: nothing we hold reports it, for any model.
| Provider | Tool calling | JSON output | Strict schema |
|---|---|---|---|
| OpenRouterOpenRouter's own listing | ✓ | ✓ | ✓ |
| Novita AIfp8Direct and through OpenRouter | ✓ | ✓ | ✓ |
Tool calling: 2 of 2 listings say yes. JSON output: 2 of 2 listings say yes. Strict schema: 2 of 2 listings say yes.
Models people weigh against Kimi K2 0905
When we formed this view
Recent changes
What moved
first indexed by our pipelineEach date is the day we first saw the change, or the day the maker announced it.
What we do not know about this model yet
- We hold no measured file for it, so all 3 sizes on this page are calculated from the parameter count.
- No independent board has scored it, so we hold no quality figures at all.
- Nothing we hold says whether an endpoint streams, so we do not show it either way.
- 1 of 2 listings does not say whether it trains on prompts, and 1 answers only through OpenRouter, not for its own listing.
- We hold no cached-input rate for any of its listings.
- We hold no batch or off-peak rate for any of its listings.
- We hold a decode speed for it, but no prompt-processing (prefill) figure, so how long the input side of a job takes is unknown to us.
Licence and identifiers
What the licence allowsCustom licence, what it allows commercially, and the identifiers you need to pull this model — its Hugging Face repo, our slug and a machine-readable card.
Licence
Custom licence
This model ships custom license terms that don't map to a known template. We haven't parsed them, so commercial use, redistribution and derivatives are unverified — review the original terms before shipping.
Identifiers
- Hugging Face
- moonshotai/Kimi-K2-Instruct-0905
- Architecture
- Mixture of experts
- Takes in, gives back
- Text in, text out
- Catalogue slug
- moonshotai-kimi-k2-0905