GPT-6 Luna vs MiMo V2.6 Flash: Benchmark Comparison
GPT-6 Luna has the higher score on 8 of 20 shared benchmarks; MiMo V2.6 Flash leads on 10.
The largest observed score gap is 10.61 pts on Terminal-Bench 4.0 , where MiMo V2.6 Flash leads.
Reported ±1 standard-error ranges overlap on 14 of 20 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-6 Luna | MiMo V2.6 Flash | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 51.22% ±1.07 | 53.23% ±1.10 | 2.02 pts | Reported ±1 SE ranges overlap |
| Harvey's Legal Agent Benchmark | 2.92% ±1.36 | 11.25% ±2.54 | 8.33 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 30.29% ±3.19 | 37.98% ±3.37 | 7.69 pts | Reported ±1 SE ranges do not overlap |
| EMB | 68.52% ±2.81 | 65.46% ±2.71 | 3.06 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 49.87% ±0.23 | 56.28% ±0.46 | 6.40 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 58.86% ±3.21 | 59.90% ±3.13 | 1.04 pts | Reported ±1 SE ranges overlap |
| MedCode | 44.69% ±2.30 | 41.06% ±2.03 | 3.63 pts | Reported ±1 SE ranges overlap |
| MedScribe | 83.71% ±1.95 | 85.28% ±1.99 | 1.57 pts | Reported ±1 SE ranges overlap |
| ProofBench v1.1 | 64.00% ±4.82 | 63.00% ±4.85 | 1.00 pts | Reported ±1 SE ranges overlap |
| BioMysteryBench | 61.48% ±2.59 | 69.26% ±0.98 | 7.78 pts | Reported ±1 SE ranges do not overlap |
| MysteryMechanism | 19.37% ±2.66 | 21.62% ±2.77 | 2.25 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench Science | 4.29% ±2.44 | 4.29% ±2.44 | 0.00 pts | Reported ±1 SE ranges overlap |
| SAGE | 48.09% ±3.38 | 43.53% ±3.43 | 4.56 pts | Reported ±1 SE ranges overlap |
| Code Migration | 42.55% ±4.42 | 40.93% ±4.35 | 1.63 pts | Reported ±1 SE ranges overlap |
| IOI | 55.56% ±8.87 | 47.67% ±4.79 | 7.89 pts | Reported ±1 SE ranges overlap |
| ProgramBench | 0.50% ±0.50 | 0.50% ±0.50 | 0.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 13.64% ±1.51 | 24.24% ±1.51 | 10.61 pts | Reported ±1 SE ranges do not overlap |
| Vibe Code Bench v1.1 | 81.65% ±3.38 | 78.96% ±4.04 | 2.69 pts | Reported ±1 SE ranges overlap |
| CyberBench v1.1 | 76.25% ±5.29 | 75.36% ±5.42 | 0.89 pts | Reported ±1 SE ranges overlap |
| Public Benefits Bench v1.1 | 57.65% ±1.28 | 67.59% ±1.22 | 9.95 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | GPT-6 Luna average | MiMo V2.6 Flash average |
|---|---|---|
| Legal | 16.60% | 24.62% |
| Finance | 59.08% | 60.55% |
| Healthcare | 64.20% | 63.17% |
| Math | 64.00% | 63.00% |
| Science | 28.38% | 31.72% |
| Education | 48.09% | 43.53% |
| Coding | 38.78% | 38.46% |
| Cyber | 76.25% | 75.36% |
| Social Mobility | 57.65% | 67.59% |
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-6 Luna cost | MiMo V2.6 Flash cost | GPT-6 Luna latency | MiMo V2.6 Flash latency |
|---|---|---|---|---|
| Vals Index | $0.43 | $0.20 | 30m34s | 58m02s |
| Harvey's Legal Agent Benchmark | $0.30 | $0.09 | 17m32s | 17m12s |
| Legal Research Bench | $0.44 | $0.07 | 42m28s | 15m20s |
| EMB | $0.14 | $0.11 | 17m21s | 25m22s |
| Finance Agent (v2) | $0.12 | $0.07 | 14m58s | 7m02s |
| Tax Agent Bench | $0.19 | $0.05 | 33m16s | 17m19s |
| MedCode | N/A | N/A | 108.77s | 93.37s |
| MedScribe | N/A | N/A | 2m50s | 74.48s |
| ProofBench v1.1 | $0.04 | $0.12 | 7m31s | 1h03m |
| BioMysteryBench | $0.05 | $0.05 | 8m44s | 28m28s |
| MysteryMechanism | $0.03 | $0.06 | 6m01s | 35m08s |
| Terminal-Bench Science | $0.24 | $0.24 | 1h16m | 3h59m |
| SAGE | N/A | N/A | 91.07s | 2m16s |
| Code Migration | $0.60 | $0.49 | 50m23s | 3h12m |
| IOI | $0.20 | $0.22 | 35m47s | 1h49m |
| ProgramBench | $0.18 | $0.47 | 28m55s | 3h22m |
| Terminal-Bench 4.0 | $0.35 | $0.22 | 32m46s | 2h25m |
| Vibe Code Bench v1.1 | $1.35 | $0.56 | 37m00s | 50m13s |
| CyberBench v1.1 | $0.13 | $0.05 | 15m50s | 28m22s |
| Public Benefits Bench v1.1 | $0.22 | $0.03 | 52m59s | 20m56s |
Results available only for GPT-6 Luna
None.
Results available only for MiMo V2.6 Flash
- SRE Bench