GPT-5.6 Luna vs MiMo V2.6 Flash: Benchmark Comparison

GPT-5.6 Luna has the higher score on 2 of 17 shared benchmarks; MiMo V2.6 Flash leads on 15.

The largest observed score gap is 14.11 pts on IOI , where GPT-5.6 Luna leads.

Reported ±1 standard-error ranges overlap on 9 of 17 shared benchmarks with comparable uncertainty data. This is not a pairwise statistical significance test.

Shared benchmark results

Scores and reported standard errors come from the latest visible Vals benchmark version available for both models.

Benchmark GPT-5.6 Luna MiMo V2.6 Flash Gap Reported uncertainty
Vals Science Index 25.30% ±0.82 31.72% ±1.27 6.42 pts Reported ±1 SE ranges do not overlap
Legal Research Bench 36.54% ±3.35 37.98% ±3.37 1.44 pts Reported ±1 SE ranges overlap
Tax Agent Bench 24.82% ±3.11 28.51% ±3.14 3.69 pts Reported ±1 SE ranges overlap
MedCode 42.39% ±2.27 41.06% ±2.03 1.33 pts Reported ±1 SE ranges overlap
MedScribe 84.39% ±2.58 85.28% ±1.99 0.88 pts Reported ±1 SE ranges overlap
ProofBench v1.1 60.00% ±4.92 63.00% ±4.85 3.00 pts Reported ±1 SE ranges overlap
BioMysteryBench 61.48% ±0.74 69.26% ±0.98 7.78 pts Reported ±1 SE ranges do not overlap
MysteryMechanism 14.41% ±2.36 21.62% ±2.77 7.21 pts Reported ±1 SE ranges do not overlap
Terminal-Bench-Science 0.00% ±0.00 4.29% ±2.44 4.29 pts Reported ±1 SE ranges do not overlap
IOI 61.78% ±11.61 47.67% ±4.79 14.11 pts Reported ±1 SE ranges overlap
ProgramBench 0.00% ±0.00 0.50% ±0.50 0.50 pts Reported ±1 SE ranges overlap
Terminal-Bench 4.0 11.62% ±1.01 24.24% ±1.51 12.63 pts Reported ±1 SE ranges do not overlap
Vibe Code Bench v1.1 77.06% ±3.07 78.96% ±4.04 1.90 pts Reported ±1 SE ranges overlap
CyberBench v1.1 73.63% ±5.56 75.36% ±5.42 1.73 pts Reported ±1 SE ranges overlap
SRE Bench 1.53% ±0.76 4.20% ±1.24 2.67 pts Reported ±1 SE ranges do not overlap
Access to Justice Benchmark 65.56% ±0.77 70.75% ±0.73 5.19 pts Reported ±1 SE ranges do not overlap
Public Benefits Bench v1.1 61.16% ±1.27 67.59% ±1.22 6.43 pts Reported ±1 SE ranges do not overlap

Performance by category

Category GPT-5.6 Luna average MiMo V2.6 Flash average
Legal 36.54% 37.98%
Finance 24.82% 28.51%
Healthcare 63.39% 63.17%
Math 60.00% 63.00%
Science 25.30% 31.72%
Coding 37.61% 37.84%
Cyber 37.58% 39.78%
Social Mobility 63.36% 69.17%

Cost and latency

Cost per test appears only where the benchmark reports it for both models. Latency is the measured completion time for that benchmark.

Benchmark GPT-5.6 Luna cost MiMo V2.6 Flash cost GPT-5.6 Luna latency MiMo V2.6 Flash latency
Vals Science Index $0.26 $0.12 50m00s 1h41m
Legal Research Bench $0.85 $0.07 39m34s 15m20s
Tax Agent Bench $0.47 $0.05 42m46s 17m19s
MedCode N/A N/A 81.28s 93.37s
MedScribe N/A N/A 116.89s 74.48s
ProofBench v1.1 $0.12 $0.12 8m34s 1h03m
BioMysteryBench $0.09 $0.05 18m31s 28m28s
MysteryMechanism $0.11 $0.06 8m13s 35m08s
Terminal-Bench-Science $0.58 $0.24 2h03m 3h59m
IOI $0.58 $0.22 1h10m 1h49m
ProgramBench $0.94 $0.47 41m34s 3h22m
Terminal-Bench 4.0 $0.73 $0.21 35m55s 2h25m
Vibe Code Bench v1.1 $0.73 $0.56 25m52s 50m13s
CyberBench v1.1 $0.40 $0.05 16m15s 28m22s
SRE Bench $3.44 $0.40 1h42m 3h22m
Access to Justice Benchmark $0.09 $0.02 5m16s 12m34s
Public Benefits Bench v1.1 $0.27 $0.03 38m40s 20m56s

Results available only for GPT-5.6 Luna

  • LegalBench
  • MortgageTax
  • TaxEval v2
  • GPQA Diamond
  • MMLU Pro
  • MMMU Pro
  • SkillsBench
  • SWE-bench
  • Financial Coaching Bench

Results available only for MiMo V2.6 Flash

  • Vals Index
  • Vals Finance Index
  • Vals Legal Index
  • HLAB-Vals-1.1
  • EMB v1.1
  • Finance Agent v2.1
  • SAGE
  • Code Migration v1.1
  • Vibe Code Bench 1-100 v1.1
Model details GPT-5.6 Luna Model details MiMo V2.6 Flash