Model specification · Moonshot AI
Kimi K2.5
moonshotai/kimi-k2.5
Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability and a self-directed agent swarm paradigm. Built on Kimi K2 with continued pretraining over approximately 15T mixed...
SpecificationsReported values
- Context window
- 262K
- Maximum output
- 262K
- Architecture
- text+image->text
- Tokenizer
- Other
- Knowledge cutoff
- —
- Moderated
- No
- Input modalities
- text image
- Output modalities
- text
Full pricingUSD where reported
| Price | Standard |
|---|---|
| Input | $0.570 / M |
| Output | $2.85 / M |
Benchmark indexesReported indexes
- Intelligence
- 35.4
- Coding
- 46.8
- Agentic
- 21.7
- Design Arena Elo
- 1217
Example workload costsStandard pricing only
| Workload | Input / output | Cache share | Estimated cost |
|---|---|---|---|
| Quick check | 10K / 1K × 1 | 0% | $0.00855 |
| Long analysis | 100K / 4K × 1 | 50% | $0.045 |
| Workday | 20K / 2K × 1,000 | 30% | $14.25 |
More from Moonshot AIAll Moonshot AI models
| Model | Context | Input / M | Output / M | Intel. |
|---|---|---|---|---|
| Kimi K3 | 1.0M | $3.00 | $15.00 | 57.1 |
| Kimi K2.6 | 262K | $0.600 | $3.41 | 44.2 |
| Kimi K2.7 Code | 262K | $0.730 | $3.50 | 41.9 |
| Kimi K2 Thinking | 262K | $0.600 | $2.50 | 17.3 |