MiMo-V2-Flash (Feb 2026)
MiMo-V2-Flash (Feb 2026) is a model from Xiaomi with a 256k-token context window. No provider we track lists it yet.
Providers serving MiMo-V2-Flash (Feb 2026)
No provider we track lists this model yet.
MiMo-V2-Flash (Feb 2026) benchmark scores
Compare every model in Bench rankingsPublished scores only, each linked to its source. Ticks are every other model's best published score.
GPQA Diamond
PhD-level scientific reasoning, pass@1, no tools
thinking on 83.8%v, vendor-reported by Xiaomi, thinking onJoint 92nd of 221 models with a score
Humanity's Last Exam (text-only)
Reasoning & knowledge, no tools
22.1%, source: Artificial AnalysisJoint 95th of 221 models with a score
τ²-Bench Telecom
Agentic tool use (telecom), with a Qwen3 model as the simulated user
93.3%, source: Artificial AnalysisRanks 26th of 160 models with a score
IFBench
Instruction following, 294 prompts, 5 runs each
71.8%, source: Artificial AnalysisRanks 34th of 158 models with a score
Terminal-Bench Hard
Agentic coding & terminal use, Artificial Analysis's 44-task hard subset
31.1%, source: Artificial AnalysisJoint 43rd of 157 models with a score
LCR
Long-context reasoning, AA-LCR v1.1
71.3%, source: Artificial AnalysisRanks 92nd of 211 models with a score
Omni Accuracy
Omniscience knowledge accuracy
19.8%, source: Artificial AnalysisRanks 111th of 216 models with a score
Non-Hallucination
1 − hallucination rate
51.6%, source: Artificial AnalysisJoint 60th of 216 models with a score
Each score is the model's best published result across its settings, named beside it. A setting above the default can use more tokens per task, which a per-token price does not show.
Show as table
| Benchmark | Score | Setting |
|---|---|---|
GPQA Diamond PhD-level scientific reasoning, pass@1, no tools | 83.8%v, vendor-reported by Xiaomi, thinking on | thinking on |
Humanity's Last Exam (text-only) Reasoning & knowledge, no tools | 22.1%, source: Artificial Analysis | Default |
τ²-Bench Telecom Agentic tool use (telecom), with a Qwen3 model as the simulated user | 93.3%, source: Artificial Analysis | Default |
IFBench Instruction following, 294 prompts, 5 runs each | 71.8%, source: Artificial Analysis | Default |
Terminal-Bench Hard Agentic coding & terminal use, Artificial Analysis's 44-task hard subset | 31.1%, source: Artificial Analysis | Default |
LCR Long-context reasoning, AA-LCR v1.1 | 71.3%, source: Artificial Analysis | Default |
Omni Accuracy Omniscience knowledge accuracy | 19.8%, source: Artificial Analysis | Default |
Non-Hallucination 1 − hallucination rate | 51.6%, source: Artificial Analysis | Default |
More from Xiaomi
- MiMo-V2.6-Flashfrom $0.14 in · 1.05M · MMMU-Pro (standard) 73.1% · 3 providers
- MiMo-V2.6-Profrom $0.43 in · 1.05M · Humanity's Last Exam (text-only) 49.4% · 3 providers
- MiMo-V2.5-Profrom $0.80 in · 1.05M · GPQA 86.6% · 1 provider
- MiMo-V2-Flash262k · GPQA 84.6%
- MiMo-V2-Omni256k · GPQA 82.8%
- MiMo-V2-Omni-0327GPQA 85.5%
- MiMo-V2.51.05M · GPQA 84.9%