Compare Models

MiMo V2.6 Flash

MiMo V2.6 Flash is an open-weight model from Xiaomi, released September 21, 2026. It ranks #17 of 44 models on the Vals Index with 53.23%. Its best result is #5 of 44 on CyberBench v1.1.

Release Date: Sep 21, 2026

Developer Xiaomi  🇨🇳
Context Window 1M
Max Output Tokens 128k
Token Costs (in/out) $0.14/0.28
Weights Open
Input Modalities

Accuracy

53.23 % ± 1.10

Cost / Test (Vals Index)

$ 0.202

Latency

58 min 2 s

Vals Index
BenchmarksAccuracyRankings
Vals Index

53.23%

±1.10
17/44
Code Migration

40.93%

±4.35
26/74
CyberBench v1.1

75.36%

±5.42
5/44
EMB

65.46%

±2.71
21/71
Finance Agent (v2)

56.28%

±0.46
17/75
Legal Research Bench

37.98%

±3.37
33/74
MedCode

41.06%

±2.03
56/105
MedScribe

85.28%

±1.99
28/107
MysteryMechanism

21.62%

±2.77
17/24
ProofBench v1.1

63.00%

±4.85
17/47
SAGE

43.53%

±3.43
48/90
Public Benefits Bench v1.1

67.59%

±1.22
12/47
SRE Bench

4.20%

±1.24
16/32
Tax Agent Bench

59.90%

±3.13
34/66
Vibe Code Bench v1.1

78.96%

±4.04
23/109
BioMysteryBench

69.26%

±0.98
13/24
Harvey's Legal Agent Benchmark

11.25%

±2.54
11/75
IOI

47.67%

±4.79
25/41
ProgramBench

0.50%

±0.50
18/62
Terminal-Bench 4.0

24.24%

±1.51
17/44
Terminal-Bench Science

4.29%

±2.44
28/38
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Xiaomi
Temperature: 1
Top P: 0.95
Top K: Default
Max Output Tokens: 128,000

Compared with

MiMo V2.6 Flash vs GPT-5.6 Terra 19 shared benchmarks MiMo V2.6 Flash vs Claude Sonnet 5 19 shared benchmarks MiMo V2.6 Flash vs GPT-5.6 Luna 21 shared benchmarks MiMo V2.6 Flash vs Gemini 3.7 Flash 18 shared benchmarks MiMo V2.6 Flash vs Grok 4.6 20 shared benchmarks MiMo V2.6 Flash vs Grok 4.7 21 shared benchmarks MiMo V2.6 Flash vs Muse Spark 1.3 14 shared benchmarks MiMo V2.6 Flash vs GPT-6 Luna 21 shared benchmarks MiMo V2.6 Flash vs Claude Fable 5 16 shared benchmarks MiMo V2.6 Flash vs Claude Opus 5.5 21 shared benchmarks