xiaomi
MiMo V2.6 Flash
MiMo V2.6 Flash is Xiaomi's efficient multimodal reasoning model for coding, automation, and everyday agent workflows. It accepts text, images, audio, and video within a 1M-token context window, with up to 128K tokens of output. Deep thinking, tool calling, JSON mode, and prompt caching make it suitable for applications that need multimodal understanding at a lower token cost.
audio inputimplicit cachingreasoningstructured outputtool usevideo inputvisiontextaudioimagevideo
Overview
- Type
- language
- Context
- 1M
- Max output
- 131K
- Input / output
- text, audio, image, video / text
API
Model identifier
xiaomi/mimo-v2.6-flashProviders and live performance
Rolling Gateway metrics. Latency is time to first token; throughput is output tokens per second.
| Provider | Status | Latency | Throughput | Uptime 15m / 1h / 1d | Input | Output | Cache read / write | Discount | Regions | ZDR | No training |
|---|---|---|---|---|---|---|---|---|---|---|---|
| boundless | Operational | 0.92s / 1.08s | 97.0 / 111.4 tok/s | 100.00% / 100.00% / 100.00% | $0.14/M | $0.28/M | $0.00/M / — | — | — | No | No |
| deepinfra | Operational | 8.63s / 29.67s | 49.5 / 54.0 tok/s | 100.00% / 99.32% / 99.73% | $0.13/M | $0.27/M | $0.00/M / — | — | — | Yes | Yes |
| gmicloud | Operational | 4.12s / 7.52s | 32.5 / 37.8 tok/s | 100.00% / 100.00% / 99.92% | $0.14/M | $0.28/M | $0.00/M / — | — | — | No | No |
| inference-net | Operational | — | — | — / — / — | $0.14/M | $0.28/M | $0.00/M / — | — | — | No | No |
| novita | Operational | 3.11s / 36.67s | 41.0 / 66.4 tok/s | 100.00% / 100.00% / 100.00% | $0.14/M | $0.28/M | $0.00/M / — | — | — | — | No |
| xiaomi | Operational | 3.48s / 11.51s | 43.0 / 49.6 tok/s | 99.84% / 99.82% / 99.78% | $0.14/M | $0.28/M | $0.00/M / — | — | — | No | No |
Pricing
input
0.00000014
output
0.00000028
input cache read
0.0000000028
varies by provider
Yes
Capabilities
audio inputimplicit cachingreasoningstructured outputtool usevideo inputvisiontextaudioimagevideo
Supported parameters
max tokens
Yes
temperature
Yes
stop
Yes
tools
Yes
tool choice
Yes
reasoning
Yes
include reasoning
Yes
response format
Yes
structured outputs
Yes
Reasoning options
options
{
"type": "toggle"
}, {
"type": "effort",
"values": [
"none",
"minimal",
"low",
"medium",
"high"
]
}
Metadata
type
language
context window
1M
max output
131K
released
Sep 21, 2026
temperature control
Yes
zero data retention
some
no training
some
api specifications
v2, v3, v4