Across 24 comparable benchmarks, MiMo-V2.6-Pro has better reported results on 20 and worse on 4. 1 further shared result is listed below, each with the reason it is not counted.
Moderate evidence24 of 25 shared results countedHow we compare
Price basis: International rates vs listed rates. The price section below shows the applicable periods.
| Fact | MiMo-V2.6-Pro | GPT-5.6 Luna |
|---|---|---|
| Context window | 1M tokens | 1.05M tokens |
| Input / 1M | $0.435 | $0.200 |
| Output / 1M | $0.870 | $1.20 |
| Blended (4:1) | $0.211 | $0.270 |
| Status | Available | Available |
| Released | September 22, 2026 | June 26, 2026 |
Blended is 4:1 input to output with 90% cache hits at listed rates β set your own mix.
Mean of 16 shared benchmarks grouped by capability on a 0β100 scale. How averages are built
MiMo-V2.6-Pro ahead on 6 of 8 counted
MiMo-V2.6-Pro ahead on 1 of 1 counted
MiMo-V2.6-Pro ahead on 5 of 5 counted
MiMo-V2.6-Pro ahead on 1 of 1 counted
No fixed-scale average β see the rows in Benchmark results
MiMo-V2.6-Pro ahead on 4 of 6 counted
No shared comparable results for Abstract reasoning, Research & search, Long context, Multimodal, Cybersecurity.
Published scores, with matching modes where available. Gaps show MiMo-V2.6-Pro relative to GPT-5.6 Luna. Reading the results
| Benchmark | MiMo-V2.6-Pro | GPT-5.6 Luna | Gap |
|---|---|---|---|
| APEX-SWERepository Engineering Β· Pass@1 | 54.2Auto Β· tools | 32.1β Max Β· tools | +22.1 pts#8 vs #47 of 61 |
| Code MigrationMaintenance & Optimization Β· Hidden-test pass rate (%) | 43.0Unspecified Reasoning Β· tools | 44.5β Max Β· tools | β1.5 pts#28 vs #23 of 79 |
| DeepSWERepository Engineering Β· Pass@1 (DeepSWE v1.1) | 71.9Thinking Β· tools | 67.2β Max Β· tools | +4.7 pts#21 vs #35 of 73 |
| IOI (Vals v2)Algorithmic Coding Β· Subtask points (%) | 39.3Unspecified Reasoning Β· tools | 61.8β Max Β· tools | β22.5 pts#36 vs #17 of 45 |
| Terminal-Bench 2.1Agentic Development Β· Accuracy | 89.9Thinking Β· tools | 84.7β Max Β· no tools | +5.2 pts#6 vs #28 of 68 |
| Terminal-Bench 4.0Agentic Development Β· Resolution rate (%) | 34.9Thinking Β· tools | 17.3β Max Β· tools | +17.6 pts#34 vs #52 of 90 |
| Terminal-Bench 4.1Agentic Development Β· Resolution rate (%) | 27.0Normal Β· tools | 17.6β Max Β· tools | +9.4 pts#27 vs #34 of 48 |
| Vibe Code Bench v1.1Code Generation & Editing Β· Pass rate (%) | 85.2Unspecified Reasoning Β· tools | 77.1β Max Β· tools | +8.2 pts#20 vs #41 of 121 |
| ProofBench v1.1 (Lean 4)Mathematics Β· Accuracy (%) | 70.0Unspecified Reasoning Β· tools | 60.0β Max Β· tools | +10.0 pts#22 vs #26 of 54 |
| CritPtScientific Reasoning Β· Score | 26.6Thinking Β· no tools | 20.6β Extra-High Β· no tools | 1.29Γ#19 vs #28 of 134 |
| MysteryMechanismScientific Reasoning Β· Accuracy (%) | 15.3Unspecified Reasoning Β· tools | 14.4β Max Β· tools | +0.9 pts#30 vs #31 of 32 |
| SciCodeScientific Computing Β· Score | 60.9Thinking Β· no tools | 53.6β Max Β· no tools | 1.14Γ#8 vs #37 of 98 |
| MedCodeClinical Workflows Β· Accuracy (%) | 45.0Unspecified Reasoning Β· tools | 42.4β Max Β· tools | +2.6 pts#37 vs #49 of 97 |
| MedScribeClinical Workflows Β· Rubric pass rate (%) | 88.3Unspecified Reasoning Β· tools | 84.4β Max Β· tools | +3.9 pts#13 vs #35 of 99 |
| APEX-AgentsTool Orchestration Β· Score | 54.4Reported best | 43.0Reported best | 1.27Γ#30 vs #44 of 61 |
| AutomationBenchOffice & Business Β· Pass RateNot counted Β· Different or unrecorded test environments β excluded from the tally | 53.1Thinking Β· toolsΒ· legacy-default | 17.1β Max Β· toolsΒ· zapier-private-heldout | β#7 vs #40 of 64 |
| EMB (Excel Modeling)Finance Β· Accuracy (%) | 62.9Unspecified Reasoning Β· tools | 67.1β Max Β· tools | β4.3 pts#32 vs #20 of 78 |
| Finance Agent v2Finance Β· Score | 57.3Unspecified Reasoning Β· tools | 55.0β Max Β· tools | 1.04Γ#14 vs #23 of 75 |
| Legal Research BenchLegal Β· Score | 47.1Unspecified Reasoning Β· tools | 36.5β Max Β· tools | 1.29Γ#16 vs #48 of 91 |
| Public Benefits Bench v1.1Service Workflows Β· Criterion pass rate (%) | 68.9Unspecified Reasoning Β· tools | 61.2β Max Β· tools | +7.8 pts#8 vs #31 of 50 |
| SAGEService Workflows Β· Accuracy (%) | 45.0Unspecified Reasoning Β· tools | 44.2β Max Β· tools | +0.8 pts#41 vs #45 of 82 |
| Tax Agent BenchFinance Β· Accuracy (%, rubric Γ citation quality) | 32.4Unspecified Reasoning Β· tools | 60.8β Max Β· tools | β28.4 pts#34 vs #17 of 74 |
| AA Intelligence Index v5.0.1Capability Indices Β· Index (0-100) | 29.9Thinking Β· tools | 23.8β Max Β· task tools | 1.26Γ#31 vs #41 of 62 |
| LiveBench 2.0Cross-capability Suites Β· Overall score (equal-weight mean of 7 suite scores, 0-100) | 74.9Thinking High Β· tools | 66.1β Max Β· task tools | 1.13Γ#19 vs #31 of 51 |
| Vals Index v2.1Capability Indices Β· GDP-weighted accuracy (%) | 55.2Reported best | 51.7β Max Β· task tools | +3.5 pts#12 vs #22 of 38 |
Repository Engineering Β· Pass@1
Maintenance & Optimization Β· Hidden-test pass rate (%)
Repository Engineering Β· Pass@1 (DeepSWE v1.1)
Algorithmic Coding Β· Subtask points (%)
Agentic Development Β· Accuracy
Agentic Development Β· Resolution rate (%)
Agentic Development Β· Resolution rate (%)
Code Generation & Editing Β· Pass rate (%)
Mathematics Β· Accuracy (%)
Scientific Reasoning Β· Score
Scientific Reasoning Β· Accuracy (%)
Scientific Computing Β· Score
Clinical Workflows Β· Accuracy (%)
Clinical Workflows Β· Rubric pass rate (%)
Tool Orchestration Β· Score
Office & Business Β· Pass Rate
Not counted Β· Different or unrecorded test environments β excluded from the tally
Finance Β· Accuracy (%)
Finance Β· Score
Legal Β· Score
Service Workflows Β· Criterion pass rate (%)
Service Workflows Β· Accuracy (%)
Finance Β· Accuracy (%, rubric Γ citation quality)
Capability Indices Β· Index (0-100)
Cross-capability Suites Β· Overall score (equal-weight mean of 7 suite scores, 0-100)
Capability Indices Β· GDP-weighted accuracy (%)
A dash means no result is available.
Time-of-day providers use peak-hour rates; fixed-price providers use standard rates. Off-peak discounts remain in the published details.
Drag the handle above the bar for cache hits, the one below it for the input : output mix Β· 5% steps
Blended price = effective input rate at 90% cache hits Γ 80% + output rate Γ 20%
MiMo-V2.6-Pro costs22%lessthan GPT-5.6 Luna per million tokens.Using international rates vs listed rates at this token mix.
MiMo-V2.6-Pro: International pricing. Xiaomi MiMo international pricing β input Xiaomi MiMo international pricing β output
GPT-5.6 Luna: Standard pricing. context_length <= 272000 OpenAI official pricing (developers.openai.com/api/docs/pricing) verified 2026-10-21; standard tier; <=272K input tokens context_length > 272000 OpenAI official pricing (developers.openai.com/api/docs/pricing) verified 2026-10-21; standard tier; >272K input tokens; start date not published, verification date used
| Price / 1M tokens | MiMo-V2.6-Pro | GPT-5.6 Luna | Ratio |
|---|---|---|---|
| Input | $0.435 | $0.20 | 2.17Γ |
| Output | $0.87 | $1.20 | 0.72Γ |
| Cached input | $0.004 | $0.02 | 0.18Γ |
| Cache write | β | $0.25 | β |
| Blended | $0.211 | $0.27 | 0.78Γ |
Estimates use the selected token mix. Cache writes, tool fees and long-context surcharges are excluded.
Reported cost per task, paired with the score from the same run. Settings and sources may differ; costs do not affect the verdict.
MiMo-V2.6-Pro relative to GPT-5.6 Luna
Summary: 1 of 5 shared benchmarks have matching recorded settings and sources. The average is a geometric mean of 1 positive-cost ratios, one equal weight per benchmark. This benchmark set can include related tests and indexes; it is not a workload cost estimate.
MiMo-V2.6-Pro 1.6Γ cheaper
Settings or sources differ
Settings or sources differ
Settings or sources differ
Settings or sources differ
MiMo-V2.6-Pro 1.6Γ cheaper
Settings or sources differ
Settings or sources differ
Settings or sources differ
Settings or sources differ
Settings or sources differ or are not recorded; interpret the comparison with care.
MiMo-V2.6-Pro
GPT-5.6 Luna
Settings or sources differ or are not recorded; interpret the comparison with care.
MiMo-V2.6-Pro
GPT-5.6 Luna
Effort, protocol, aggregation and source match; other run conditions may differ.
MiMo-V2.6-Pro
GPT-5.6 Luna
Settings or sources differ or are not recorded; interpret the comparison with care.
MiMo-V2.6-Pro
GPT-5.6 Luna
Settings or sources differ or are not recorded; interpret the comparison with care.
MiMo-V2.6-Pro
GPT-5.6 Luna
| Benchmark | MiMo-V2.6-Pro | GPT-5.6 Luna |
|---|---|---|
| AA-LCR | Not reported | |
| Agents' Last Exam v1.0 | Not reported | 2 more settings |
| ARC-AGI-1 | Not reported | 2 more settings$0.06 / task Average Β· Score (%): 56.5 Thinking Medium (With Tools) |
| ARC-AGI-2 | Not reported | $0.29 / task Average Β· ARC-AGI-2 Score: 29.3 Thinking High (With Tools) $0.07 / task Average Β· ARC-AGI-2 Score: 5.1 Thinking Low (With Tools) 2 more settings$0.11 / task Average Β· ARC-AGI-2 Score: 7.4 Thinking Medium (With Tools) $0.54 / task Average Β· ARC-AGI-2 Score: 47.6 Extra-High (With Tools) |
| AutomationBench | Not reported | $0.02 / task Average Β· Pass Rate: 0.15 Reported best (effort unspecified) Β· zapier-private-heldout $0.05 / task Average Β· Pass Rate: 9.1 Thinking High (With Tools) Β· zapier-private-heldout 3 more settings$0.01 / task Average Β· Pass Rate: 1.8 Thinking Low (With Tools) Β· zapier-private-heldout $0.02 / task Average Β· Pass Rate: 4.3 Thinking Medium (With Tools) Β· zapier-private-heldout $0.06 / task Average Β· Pass Rate: 12.9 Extra-High (With Tools) Β· zapier-private-heldout |
| Code Migration | $0.6916 / task Average Β· Hidden-test pass rate (%): 43.0 Unspecified Reasoning (With Tools) | Not reported |
| CritPt | Not reported | |
| CursorBench 4.0 | Not reported | $1.03 / task Average Β· Task score: 35.9 Max (With Tools) Β· cursorbench-4-0-cursor-agent $0.25 / task Average Β· Task score: 29.4 Thinking High (With Tools) Β· cursorbench-4-0-cursor-agent $0.03 / task Average Β· Task score: 16.0 Thinking Low (With Tools) Β· cursorbench-4-0-cursor-agent 2 more settings$0.08 / task Average Β· Task score: 22.2 Thinking Medium (With Tools) Β· cursorbench-4-0-cursor-agent $0.44 / task Average Β· Task score: 33.0 Extra-High (With Tools) Β· cursorbench-4-0-cursor-agent |
| DeepSWE | Not reported | 2 more settings |
| EMB (Excel Modeling) | Not reported | |
| Finance Agent v2 | Not reported | |
| GDPval-AA v2 | Not reported | |
| IOI (Vals v2) | Not reported | |
| Legal Research Bench | Not reported | |
| MedCode | Not reported | |
| MedScribe | $0.00988 / task Average Β· Rubric pass rate (%): 88.3 Unspecified Reasoning (With Tools) | Not reported |
| MysteryMechanism | Not reported | |
| ProofBench v1.1 (Lean 4) | Not reported | |
| Public Benefits Bench v1.1 | $0.0819 / task Average Β· Criterion pass rate (%): 68.9 Unspecified Reasoning (With Tools) | Not reported |
| SAGE | Not reported | |
| SciCode | Not reported | |
| SWE-sweep | Not reported | |
| Tax Agent Bench | $0.1343 / task Average Β· Accuracy (%, rubric Γ citation quality): 32.4 Unspecified Reasoning (With Tools) | Not reported |
| Vals Index v2.2 | $0.4347 / task Average Β· GDP-weighted accuracy (%): 57.8 Reported best (effort unspecified) | Not reported |
| Vibe Code Bench 1-100 | Not reported | |
| Vibe Code Bench v1.1 | Not reported | |
| WeirdML v3 | Not reported | $0.7354 / task Average Β· Official score (0-1; higher is better): 0.052 Extra-High (With Tools) |
| Attribute | MiMo-V2.6-Pro | GPT-5.6 Luna |
|---|---|---|
| Availability | Available | Available |
| Provider | Xiaomi | OpenAI |
| Model type | Multimodal model | Reasoning model |
| First release | September 22, 2026 | June 26, 2026 |
| Context window | 1M tokens | 1.05M tokens |
| Max output | β | 128K tokens |
| Total parameters | 1.0T | β |
| Active parameters | 42B | β |
| Architecture | Mixture of Experts (MoE) | Undisclosed |
| Documented capabilities | Reasoning, Multilingual, Coding, Tool use, Function calling, Computer use, Long context | Reasoning, Multilingual, Tool use, Function calling, Coding, Long context |
| Knowledge cutoff | β | 2026-02-16 |
| Input modalities | Text, Image, Audio, Video | Text, Image |
| Output modalities | Text | Text |
| Weights | Open weights | Closed / API only |
| Licence | MIT License | β |
| Commercial use | Commercial use allowed | Proprietary |
These are reported-score comparisons, not controlled head-to-head tests.
What people found comparing these models, or using either one. Comments also appear on the models' own pages.
Used MiMo-V2.6-Pro or GPT-5.6 Luna? Sign in with Google to share what worked, what did not, and which API or setup you tested.
No community comments on either model yet.
Alternatives with the most benchmark overlap for this matchup.