xiaomi · Open weights
MiMo V2.5 Thinking
MiMo V2.5 with Xiaomi thinking enabled. It supports deep reasoning, tool calling, structured outputs, and web search with up to 1M context.
- Reasoning
- Tools
- Vision
- Structured output
- Audio input
- Video input
- Input
- Text, Images, Video, Audio
- Output
- Text
- Reasoning effort
- none · high
- Discovery tier
- Open weights
Pricing
USD per 1M tokens- Input
- $0.14
- Output
- $0.28
- Cache read
- $0.0028
- Cache write
- $0.00
Reference prices as published by the model provider. API example
Specifications
- Context window
- 1M
- Max output
- 131.1K
- Average output (7 days)
- 1.1K tokens
- Longer than 73% of 468 models
Side by side
MiMo V2.5 Thinking compared
Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.
| MiMo V2.5 ThinkingxiaomiThis model | Kimi K2 ThinkingMoonshotRemove | |
|---|---|---|
| Provider | xiaomi | Moonshot |
| Discovery tier | Open weights | Open weights |
| Added | Jun 3, 2026 | Nov 6, 2025 |
| Limits | ||
| Context window | 1.05MBest | 262.14K |
| Max output | 131.07KBest | 98.3K |
| Average output | 1.1K tokens | 1.9K tokens |
| Pricing, USD per 1M tokens | ||
| Input | $0.14Best | $0.60 |
| Output | $0.28Best | $2.50 |
| Cache read | $0.0028Best | $0.15 |
| Cache write | $0.00 | — |
| Performance | ||
| Throughput | 28.6 tok/s | 78.6 tok/sBest |
| Time to first token | 4.1 sBest | 4.6 s |
| Request success (24h) | 100.0%Best | 99.6% |
| Benchmark indexes | ||
| Intelligence Index | 21.8 | 22.0Best |
| Coding Index | 56.8 | — |
| Math Index | — | 94.7 |
| Capabilities | ||
| Reasoning | Yes | Yes |
| Tools | Yes | Yes |
| Vision | Yes | No |
| Structured output | Yes | Yes |
| Audio input | Yes | No |
| Video input | Yes | No |
| Reasoning effort | none, high | — |
Public evaluations
Benchmarks
Independent evaluation results, with how each score ranks among the models in the Noumen catalog.
- MiMo V2.5 Thinking
- Kimi K2 Thinking
Intelligence Index
- MiMo V2.5 Thinking21.8
- Kimi K2 Thinking22.0
MiMo V2.5 Thinking: Better than 52% of 326 scored models
Indexes
Coding Index
Composite of the coding evaluations
- MiMo V2.5 Thinking: 56.8
- Kimi K2 Thinking: —
Math Index
Composite of the math evaluations
- MiMo V2.5 Thinking: —
- Kimi K2 Thinking: 94.7
Agentic work
AutomationBench-AA
Workflow automation, penalized for guardrail violations
- MiMo V2.5 Thinking: 18.4%
- Kimi K2 Thinking: —
AutomationBench-AA tasks completed
Workflows finished end to end without a guardrail violation
- MiMo V2.5 Thinking: 3.3%
- Kimi K2 Thinking: —
AA-Briefcasev1.1
Agentic knowledge work, rated head to head
- MiMo V2.5 Thinking: 754 Elo
- Kimi K2 Thinking: —
GDPval-AAv2.1
Economically valuable occupational tasks, rated head to head
- MiMo V2.5 Thinking: 996 Elo
- Kimi K2 Thinking: —
Document reasoning
GDP.pdf
Professional reasoning over PDF documents, all-pass rate
- MiMo V2.5 Thinking: 4.0%
- Kimi K2 Thinking: —
AA-LCRv1.1
Reasoning across long documents
- MiMo V2.5 Thinking: 73.0%Top
- Kimi K2 Thinking: 72.0%
Reasoning and knowledge
Humanity’s Last Exam
Expert-level questions across academic fields
- MiMo V2.5 Thinking: 27.2%Top
- Kimi K2 Thinking: 23.8%
GPQA Diamond
Graduate-level science questions
- MiMo V2.5 Thinking: 84.9%Top
- Kimi K2 Thinking: 83.8%
MMLU-Pro
Multi-subject knowledge and reasoning
- MiMo V2.5 Thinking: —
- Kimi K2 Thinking: 84.8%
IFBench
Following precise output instructions
- MiMo V2.5 Thinking: 67.1%
- Kimi K2 Thinking: 68.1%Top
Coding
Terminal-Benchv4.0
Practical software engineering in a terminal
- MiMo V2.5 Thinking: 0.0%
- Kimi K2 Thinking: —
SciCode
Scientific computing in Python
- MiMo V2.5 Thinking: 43.9%
- Kimi K2 Thinking: —
LiveCodeBench
Recent competitive programming problems
- MiMo V2.5 Thinking: —
- Kimi K2 Thinking: 85.3%
Math
AIME 2025
Competition mathematics, 2025 problems
- MiMo V2.5 Thinking: —
- Kimi K2 Thinking: 94.7%
Legacy evaluations
Terminal-Bench Hard
Earlier terminal task suite
- MiMo V2.5 Thinking: 41.7%Top
- Kimi K2 Thinking: 31.1%
τ²-Bench Telecom
Tool use in a customer-support simulation
- MiMo V2.5 Thinking: 90.6%
- Kimi K2 Thinking: 93.0%Top
AA-LCR (unversioned)
Long-context reasoning, earlier grading
- MiMo V2.5 Thinking: 73.0%Top
- Kimi K2 Thinking: 72.0%
Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis
Live routing data
Performance and reliability
Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.
- Throughput
- 28.6 tok/s
- Faster than 20% of 254 models
- Time to first token
- 4.1 s
- Quicker than 45% of 254 models
- Prompt cache hits
- 83.4%
- About $0.0256 per 1M input tokens with caching
- Request success (24h)
- 100.0%
- Reliable
Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.
- Healthy (98% or more)
- Degraded (90–98%)
- Down (under 90%)
- No requests
Developers
Use it through the API
Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.
curl https://www.noumen.xyz/api/v1/chat/completions \
-H "Authorization: Bearer $NOUMEN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "xiaomi/mimo-v2.5:thinking",
"reasoning_effort": "none",
"messages": [{ "role": "user", "content": "Explain this stack trace." }]
}'import os
from openai import OpenAI
client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
model="xiaomi/mimo-v2.5:thinking",
messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)xiaomi
Related models
Other xiaomi models, starting with the same family.