// Moonshot (Kimi)
Kimi K3
A Moonshot (Kimi) model. 11 prices across 11 channels side by side, each with its source and check date – plus specs, licence, speed and benchmarks.
- Official in/out per 1M
- ¥20 / ¥100
- Lowest (Chutes)
- $1.85 / $9.5
- Context
- 1.05M
- AA index
- 43.6
Prices checked 2026-10-22Official sources linked on every rowNo estimatesCheapest LLM APIs →
01Specs at a glance
Official values first; gaps filled from OpenRouter’s API where the vendor page is silent.
| Vendor | Moonshot (Kimi) |
|---|---|
| Context window | 1.05M (1,048,576 tokens) |
| Max output | 944K |
| Input | Text · Image · Video |
| Output | Text |
| Tool calling / JSON | Yes / Yes |
| Weights & licence | Open weights · kimi-k3 (custom) · Hugging Face |
| Parameters | 2.78T (from weight files) |
| Released | Jul 16, 2026 |
| Model IDs | kimi-k3 · moonshotai/kimi-k3 · kimi-k3@bailian |
02Prices on 11 channels
Per 1M tokens, in the currency each source publishes (1 USD = 6.7 CNY for the blended column). Sorted: official first, then cheapest.
| Channel | Source | Input | Output | Cache hit | Blended $ | Context | Notes |
|---|---|---|---|---|---|---|---|
| Official Vendor’s China site, billed in CNY | Official 10-10 | ¥20 | ¥100 | ¥2 | 5.97 | 1.05M | cache write 20 (5min) / 40 (1h) |
| Chutes · mxfp4 Via OpenRouter / provider listed there | OpenRouter 10-22 | $1.85 | $9.5 | $0.185 | 3.76 | 1.05M | Chutes price as listed on OpenRouter (pass-through; quantization=mxfp4; discount field=0); not checked against Chutes's own page |
| Together Inference platform (own price page) | Platform 10-10 | $2.7 | $13.5 | $0.27 | 5.4 | 1.05M | Together AI own pricing page; promo until 2026-10-11; list 3.00/15.00 |
| DeepInfra Inference platform (own price page) | Platform 10-10 | $2.85 | $14.25 | $0.285 | 5.7 | 1.05M | DeepInfra own pricing page |
| 阿里云百炼 Inference platform (own price page) | Platform 10-10 | ¥20 | ¥100 | — | 5.97 | — | Third-party model hosted on 阿里云百炼 (Beijing) |
| Fireworks Inference platform (own price page) | Platform 10-10 | $3 | $15 | $0.3 | 6 | 1.05M | Batch −50%Fireworks AI own pricing page; batch -50%; priority 1.25x |
| Parasail · fp4 Via OpenRouter / provider listed there | OpenRouter 10-22 | $3 | $15 | $0.3 | 6 | 1.05M | Parasail price as listed on OpenRouter (pass-through; quantization=fp4; discount field=0); not checked against Parasail's own page |
| BaseTen · fp8 Via OpenRouter / provider listed there | OpenRouter 10-22 | $3 | $15 | $0.3 | 6 | 1.05M | BaseTen price as listed on OpenRouter (pass-through; quantization=fp8; discount field=0); not checked against BaseTen's own page |
| Amazon Bedrock · unknown Via OpenRouter / provider listed there | OpenRouter 10-22 | $3 | $15 | $0.3 | 6 | 1.05M | Amazon Bedrock price as listed on OpenRouter (pass-through; quantization=unknown); Amazon Bedrock; promoted from endpoint candidates 2026-10-14; not checked against the provider's own page |
| Moonshot AI (via OpenRouter) · mxfp4 Via OpenRouter / provider listed there | OpenRouter 10-22 | $3 | $15 | $0.3 | 6 | 1.05M | Moonshot AI price as listed on OpenRouter (pass-through; quantization=mxfp4); Moonshot AI (maker); promoted from endpoint candidates 2026-10-14; not checked against the provider's own page |
| Alibaba Cloud · unknown Via OpenRouter / provider listed there | OpenRouter 10-22 | $3.45 | $17.25 | $0.345 | 6.9 | 1.05M | Alibaba price as listed on OpenRouter (pass-through; quantization=unknown); Alibaba Cloud; promoted from endpoint candidates 2026-10-14; not checked against the provider's own page |
“Official” = vendor pricing page; “Platform” = the platform’s own pricing page; “OpenRouter” = price listed on OpenRouter, not checked on the platform’s site. Click a source tag to open the original page.
03Where you can use it
Derived only from the price sources above. Country restrictions are set by each vendor – check their terms before you sign up.
- Vendor’s China site, billed in CNY
Official - Via OpenRouter / provider listed there
Chutes · Parasail · BaseTen · Amazon Bedrock · Moonshot AI (via OpenRouter) · Alibaba Cloud - Inference platform (own price page)
Together · DeepInfra · 阿里云百炼 · Fireworks - Self-host the open weights
GPU / VRAM sizing and build plans→ - Free option
Kimi-K3→
04Performance
Speed and AA index from Artificial Analysis public leaderboards (not our own tests); benchmark scores from our benchmark library, best result per benchmark.
- AA indexKimi K3 (Max)
- 43.6
- Output speed (median)
- 39 tok/s
- Time to first token
- 3.9s
- AA index per $ (blended)
- 7.3
| Benchmark | Best score | Source | Date |
|---|---|---|---|
| GPQA Diamond | 93.1% | independent | 2026-07-16 |
| AIME (OTIS Mock AIME 2024–2025) | 97.2% | independent | 2026-07-16 |
| τ²-bench | 37.1% | official | 2026-07-24 |
05Price changes we recorded
All price changes →From our daily re-checks of the sources above.
- 2026-10-19 · Parasail · cache read0.26 → 0.3 USD (+15.4%)
- 2026-10-19 · Parasail · output13 → 15 USD (+15.4%)
- 2026-10-19 · Parasail · input2.6 → 3 USD (+15.4%)
- 2026-10-14 · Alibaba Cloud · new listing
- 2026-10-14 · Moonshot AI (via OpenRouter) · new listing
- 2026-10-14 · Amazon Bedrock · new listing
- 2026-10-13 · Chutes · cache read0.3 → 0.185 USD (-38.3%)
- 2026-10-13 · Chutes · output15 → 9.5 USD (-36.7%)
- 2026-10-13 · Chutes · input3 → 1.85 USD (-38.3%)
06More from Moonshot (Kimi)
FAQ
How much does the Kimi K3 API cost?
On the official pricing page: ¥20 / ¥100 per 1M input / output tokens, cached input ¥2 (checked 2026-10-10). The lowest price we found is $1.85 / $9.5 on Chutes.
What is the context window of Kimi K3?
1.05M tokens of context, up to 944K output tokens. Platforms may cap it lower – see the table.
Is Kimi K3 open source?
Yes, the weights are published on Hugging Face (moonshotai/Kimi-K3). Licence: kimi-k3 (custom).