Roboflow

Xiaomi: MiMo V2.6 Pro

MiMo V2.6 Pro Overview

MiMo V2.6 Pro is the flagship omni-modal foundation model in Xiaomi's MiMo V2.6 series, released as open weights alongside a Flash variant and a 9B distillation of Qwen3.5. It uses a sparse mixture-of-experts transformer with 1.02 trillion total parameters and roughly 42 billion activated per token, paired with a hybrid attention design that interleaves sliding-window and global attention layers to support a context window of about one million tokens. Dedicated encoders handle non-text inputs, including a vision encoder of roughly 681 million parameters and an audio tokenizer stack, so the model accepts text, images, video, and audio and returns text.

Post-training centers on large-scale reinforcement learning across thousands of interactive environments, combined with agentic grading, self-correction cold start, and a multi-prefix multi-teacher on-policy distillation stage that extends behavior to tasks that are hard to verify automatically. The resulting model targets long-horizon agentic work such as software engineering, terminal and computer-use operation, tool calling, cybersecurity analysis, and visual coding, and it reports gains over the prior MiMo generation on SWE-bench Verified, Terminal Bench, and internal visual coding and cyber benchmarks.

MiMo V2.6 Pro Interactive Demo

Model settings

Thinking level

Max output tokens

Default 65,536 · max 65,536

Sign in to adjust thinking and output length per run.

Results appear here. Add an image or pick an example to run MiMo V2.6 Pro.

MiMo V2.6 Pro Details & Performance

Details

Resources

Vision Tasks

CaptioningChart Question AnsweringClassificationDocument Question AnsweringImage TaggingMulti-Label ClassificationOCRVision LanguageVisual Question Answering

Features

Foundation VisionLLMs with Vision CapabilitiesMultimodal Vision

Usage

Past 30 Days

Performance

Avg. Latency

MiMo V2.6 Pro Vision Evals

Vision Evals is Roboflow's ground-truth benchmark: every model runs the same real-world samples across six vision tasks, and answers are scored against ground truth.

Evals updated September 22, 2026Pricing updated September 23, 2026

Overall score#49 of 59
62.5%
Avg cost / sample#15 of 59
$0.0008
Avg speed / sample#26 of 59
8.47s
Avg tokens / sample
1.7K

Strengths and weaknesses

MiMo V2.6 Pro averages 62.5% across the six Vision Evals tasks, ranking #49 of 59 models overall.

Its weakest relative showing is Identification, ranking #52 of 59 at 76.0%.

At $0.0008 per sample it is the 15th cheapest of the 59 benchmarked models, and its average inference time of 8.5s per sample makes it the 26th fastest.

Performance profile

Field medianMiMo V2.6 Pro

Field medians: Object Detection 53.9%, Counting 59.5%, Identification 84.4%, OCR 88.7%, Data Extraction 84.5%, Reasoning 54.1%.

Results by task

TaskScoreField (0 to 100)RankCost / sampleSpeed
Object Detection (low)
42.0%
±1.1, Mean of 3 runs, range 40.9 to 43.1
#40 of 59$0.001415.50s
Object Detection (high)
46.7%
±0.8, Mean of 3 runs, range 45.7 to 47.3
#19 of 25$0.003045.14s
Counting (low)
50.0%
±2.0, Mean of 3 runs, range 48.6 to 52.7
#43 of 59$0.00053.47s
Counting (high)
59.0%
±5.4, Mean of 3 runs, range 52.7 to 63.5
#23 of 25$0.001642.05s
Identification (low)
76.0%
±1.6, Mean of 3 runs, range 75.0 to 78.1
#52 of 59$0.00043.33s
Identification (high)
78.1%
±4.7, Mean of 3 runs, range 71.9 to 81.3
#25 of 25$0.001228.01s
OCR (low)
90.7%
±1.7, Mean of 3 runs, range 88.5 to 91.9
#21 of 59$0.00088.11s
OCR (high)
87.5%
±2.7, Mean of 3 runs, range 85.3 to 90.6
#22 of 25$0.0048101.54s
Data Extraction (low)
81.1%
±0.5, Mean of 3 runs, range 80.4 to 81.4
#39 of 59$0.00053.26s
Data Extraction (high)
80.4%
±1.5, Mean of 3 runs, range 79.4 to 82.5
#23 of 25$0.001325.71s
Reasoning (low)
35.1%
±2.6, Mean of 3 runs, range 32.5 to 37.8
#45 of 59$0.00053.82s
Reasoning (high)
55.9%
±2.3, Mean of 3 runs, range 54.3 to 58.9
#39 of 45$0.002668.29s
  • Thinking longer helps: 4.7 points higher on object detection at high effort for 2.2x the cost and 2.9x the latency.
  • Thinking longer helps: 9 points higher on counting at high effort for 3.3x the cost and 12.1x the latency.
  • Thinking longer helps: 2.1 points higher on identification at high effort for 2.8x the cost and 8.4x the latency.
  • Thinking longer does not help: 3.3 points lower on ocr at high effort for 6.3x the cost and 12.5x the latency.
  • Thinking longer does not help: 0.7 points lower on data extraction at high effort for 2.6x the cost and 7.9x the latency.
  • Thinking longer helps: 20.8 points higher on reasoning at high effort for 5.5x the cost and 17.9x the latency.

Price vs. performance

Score vs. cost

Overall benchmark score against estimated cost per sample, on a log scale. Upper-left is the sweet spot: high quality at low cost.

58 models on the current benchmark · scores and efficiency pooled across all six tasks at low effort · MiMo V2.6 Pro highlighted

MiMo V2.6 Pro scores are the mean of 3 runs per task at both low and high effort · Methodology

View all Vision Evals →

MiMo V2.6 Pro Pricing

MiMo V2.6 Pro costs $0.435 per 1M input tokens and $0.870 per 1M output tokens.

Input$0.435 / 1M tokens
Output$0.870 / 1M tokens
Cached input$0.004 / 1M tokens

Pricing updated Sep 23, 2026

MiMo V2.6 Pro License

MIT · Permissive license

MiMo V2.6 Pro is released under MIT, a permissive license. The MiMo V2.6 Pro license lets you use, modify, and sell work built on the model, with the copyright notice as the only real obligation and no requirement to open-source related code changes.

Commercial use
Permitted with no separate commercial license. No usage caps, revenue thresholds, or field-of-use limits apply to MiMo V2.6 Pro.
Modification
Permitted. You can fine-tune or rewrite MiMo V2.6 Pro and keep the result closed-source.
Redistribution
Permitted. Include the original copyright and permission notice in copies or substantial portions of the work.

MIT grants no explicit patent license and disclaims all warranties. If patent exposure is a concern for your deployment, review it with counsel before launch.

Read the full MIT license ↗

Do I need a commercial license for MiMo V2.6 Pro?

No commercial license is needed for MiMo V2.6 Pro: permissive terms let you keep related code private while deploying commercially.

Do not hesitate to reach out with questions for your commercial project — our team will help you start solving business problems on the first call. See Roboflow commercial licensing for the models included in each plan.

Talk to sales

This model is released under the MIT License, a short and permissive open-source license that allows commercial use, modification, and redistribution.

Yes. Under the terms of the MIT license, you can freely use this model for commercial purposes. You must retain the copyright notice and license text when redistributing.

License information is provided as a guide and is not legal advice.

Frequently Asked Questions About MiMo V2.6 Pro Vision

Yes. MiMo V2.6 Pro accepts image input and handles OCR, data extraction, object counting, identification, visual reasoning, and object detection. On Roboflow's Vision Evals its strongest task is OCR at 90.7% (#21 of 59 at low effort). You can test it on your own image in the demo above.

Yes. its transcriptions match the ground truth 90.7% on average (#21 of 59 at low effort) on Vision Evals OCR. Pulling specific fields out of documents (data extraction) scores 81.1%.

Not its strength. On Vision Evals, MiMo V2.6 Pro scores 42% mAP@50 on object detection (#40 of 59 at low effort) and 50% judge-graded accuracy on object counting. For production counting or precise localization, pairing it with a specialized detector like RF-DETR or your own trained model in a Roboflow Workflow is usually more reliable: detect the objects, then count the detections.

On our benchmark's task mix, MiMo V2.6 Pro averages $0.0008 per sample at $0.43 per 1M input and $0.87 per 1M output tokens (#15 of 59 on cost), with an average speed of 8.5s per sample across the benchmark. Actual cost depends on your images and prompts.

On the overall Vision Evals ranking, MiMo V2.6 Pro sits #49 of 59 at 62.5%, just behind Qwen3.6 27B (62.7%) and just ahead of Qwen3.7 Flash (61.5%). See the full side-by-side: MiMo V2.6 Pro vs Qwen3.6 27B.