Abacus
Kimi K2.5
Kimi multimodal agent model for visual understanding, coding, and planning
Cheapest input
$0.35
$/MTok · models.dev
Cheapest output
$1.70
$/MTok
Context
262K
tokens
Offers
20
2 sources
Offers
Same model, different listings. First-party, router, and aggregator rows stay separate so you can see the spread instead of a single blended number. Free/$0 listings are hidden by default (5 omitted). Show free/$0 offers.
| Source | Provider | Input | Output | Cache read | Context |
|---|---|---|---|---|---|
| models.devcheapestkimi-k2.5 | CrofAI | $0.35 | $1.70 | $0.07 | 262K |
| models.devkimi-k2.5 | LLM Gateway | $0.405 | $1.98 | $0.225 | 262K |
| models.devkimi-k2.5 | DigitalOcean | $0.5 | $2.70 | $0 | 262K |
| models.devkimi-k2.5 | Auriko | $0.5 | $2.80 | $0 | 262K |
| models.devkimi-k2.5 | Cortecs | $0.55 | $2.76 | $0 | 256K |
| models.devkimi-k2-5 | Venice AI | $0.56 | $3.50 | $0.22 | 256K |
| models.devkimi-k2.5 | Charm Hyper | $0.564 | $2.82 | $0 | 262K |
| models.devkimi-k2.5 | Alibaba / Qwen | $0.574 | $2.41 | $0 | 262K |
| models.devkimi-k2.5 | Abacus | $0.6 | $3 | $0 | 262K |
| models.devkimi-k2.5 | AIHubMix | $0.6 | $3 | $0.1 | 262K |
| models.devkimi/kimi-k2.5 | Alibaba / Qwen | $0.6 | $3 | $0.1 | 262K |
| models.devkimi-k2.5 | Azure | $0.6 | $3 | $0 | 262K |
| models.devkimi-k2.5 | Azure | $0.6 | $3 | $0 | 262K |
| models.devkimi-k2.5 | DaoXE | $0.6 | $3 | $0.1 | 262K |
| models.devkimi-k2.5 | FrogBot | $0.6 | $3 | $0.1 | 256K |
| models.devkimi-k2.5 | Moonshot AI (China) | $0.6 | $3 | $0.1 | 262K |
| models.devkimi-k2.5 | OpenCode Go | $0.6 | $3 | $0.1 | 262K |
| models.devkimi-k2.5 | OpenCode Zen | $0.6 | $3 | $0.08 | 262K |
| models.devkimi/kimi-k2.5 | OrcaRouter | $0.6 | $3 | $0.1 | 262K |
| LiteLLMKimi-K2.5 | — | $0.6 | $3 | $0.06 | — |
Capabilities
- Reasoning
- Tool calling
- Structured output
- Attachments
- Open weights
- In: text
- In: image
- In: video
- Out: text
Knowledge cutoff: 2025-01Released: 2026-01Family: kimi-k2
Scores
Sparse coverage on purpose. Missing a score means we have not linked one yet, not that the model is unranked. Full boards on /benchmarks.
- AA intelligence35.4
- AA coding46.8
- AA agentic21.7
Sources: Arena AI, Artificial Analysis. See /benchmarks.