Head-to-head comparison

MiMo-V2.5 vs. MiMo-V2.5-Pro

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series: 101
distinct benchmarks: 19
latest retrieval: 2026-09-04

First modelSecond model

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for MiMo-V2.5 and MiMo-V2.5-Pro
Attribute	MiMo-V2.5	MiMo-V2.5-Pro
Provider	Xiaomi	Xiaomi
Released	Apr 23, 2026	Apr 23, 2026
Availability	Open Source	Open Source
Model type	Open Weights	Open Weights
Parameters	Not published	1T, 42B active
Architecture	Mixture of Experts	Mixture of Experts
Context window	1,000,000 tokens	1,000,000 tokens
Knowledge cutoff	Not published	Not published
Technical sources	Model, Context	Model

Pricing

API pricing compared

The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.

API pricing for MiMo-V2.5 and MiMo-V2.5-Pro
Price type	MiMo-V2.5	MiMo-V2.5-Pro
API input	$0.14	$0.43
API output	$0.28	$0.87
Cache read	$0.	$0.
Cache write (5 min.)	Not listed	Not listed
Cache write (1 hr.)	Not listed	Not listed
Price verified	07/31/2026 Source	07/31/2026 Source

MiMo-V2.5MiMo-V2.5-Pro

API inputUSD per 1M tokens, base tier

MiMo-V2.5$0.14

MiMo-V2.5-Pro$0.435

API outputUSD per 1M tokens, base tier

MiMo-V2.5$0.28

MiMo-V2.5-Pro$0.87

Cache readUSD per 1M tokens, base tier

MiMo-V2.5$0.003

MiMo-V2.5-Pro$0.004

Key to the marks:better valuebetter, but not conclusivetie

Pricing by provider

Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.

Provider pricing for MiMo-V2.5 and MiMo-V2.5-Pro
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
MiMo-V2.5	Xiaomi	Direct from developerStandard API	$0.14	$0.28	Source07/31/2026
	DeepInfra	Via OpenRouterdeepinfra/fp8	$0.13	$0.65	Source09/07/2026
	GMICloud	Via OpenRoutergmicloud/fp8	Lowest listed price: $0.119	Lowest listed price: $0.238	Source09/07/2026
	Novita	Via OpenRouternovita/fp8	$0.168	$0.336	Source09/07/2026
	StreamLake	Via OpenRouterstreamlake	$0.168	$0.336	Source09/07/2026
	Xiaomi	Via OpenRouterxiaomi/fp8	$0.14	$0.28	Source09/07/2026
MiMo-V2.5-Pro	Xiaomi	Direct from developerStandard API	$0.435	$0.87	Source07/31/2026
	AtlasCloud	Via OpenRouteratlas-cloud/fp8	$0.435	$0.87	Source09/07/2026
	DeepInfra	Via OpenRouterdeepinfra/fp8	$0.39	$1.17	Source09/07/2026
	DigitalOcean	Via OpenRouterdigitalocean	$0.4	$1.5	Source09/07/2026
	GMICloud	Via OpenRoutergmicloud/bf16	Lowest listed price: $0.3045	Lowest listed price: $0.609	Source09/07/2026
	Novita	Via OpenRouternovita	$0.4802	$0.9605	Source09/07/2026
	StreamLake	Via OpenRouterstreamlake	$0.522	$1.044	Source09/07/2026
	Xiaomi	Via OpenRouterxiaomi/fp8	$0.435	$0.87	Source09/07/2026

Performance

Shared benchmarks

Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.

Capability profile from matched benchmarks

Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.

MiMo-V2.5

MiMo-V2.5-Pro

Coding: 89.2 / 96.1 (5 benchmark families)

Reasoning: 98.3 / 100 (2 benchmark families)

Mathematics: 85.1 / 100 (2 benchmark families)

Knowledge: 97.8 / 100 (2 benchmark families)

Finance: 95.5 / 100 (5 benchmark families)

Legal: 83.1 / 98.9 (4 benchmark families)

Agentic tasks: 97.9 / 90.4 (1 benchmark family)

Health: 93.9 / 100 (3 benchmark families)

101 matched rows from 19 benchmark families were considered. The radar shows 8 of 10 comparable categories.

MiMo-V2.5MiMo-V2.5-Pro

Code MigrationOverall. Higher is better. No documented setup difference

MiMo-V2.514.228 %

MiMo-V2.5-Pro21.557 %

CorpFin v2Overall. Higher is better. No documented setup difference

MiMo-V2.559.907 %

MiMo-V2.5-Pro61.422 %

EMBOverall. Higher is better. No documented setup difference

MiMo-V2.555.092 %

MiMo-V2.5-Pro55.228 %

Finance Agent (v2)Overall. Higher is better. No documented setup difference

MiMo-V2.536.728 %

MiMo-V2.5-Pro41.502 %

GPQA DiamondOverall. Higher is better. No documented setup difference

MiMo-V2.581.566 %

MiMo-V2.5-Pro82.576 %

Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. No documented setup difference

MiMo-V2.51.667 %

MiMo-V2.5-Pro2.083 %

Legal Research BenchOverall · All-pass. Higher is better. No documented setup difference

MiMo-V2.59.135 %

MiMo-V2.5-Pro15.865 %

LegalBenchOverall. Higher is better. No documented setup difference

MiMo-V2.578.886 %

MiMo-V2.5-Pro77.129 %

LiveCodeBench v6Overall. Higher is better. No documented setup difference

MiMo-V2.581.512 %

MiMo-V2.5-Pro81.354 %

LMArena Text, Style ControlOverall. Higher is better. No documented setup difference

MiMo-V2.51,433.798

MiMo-V2.5-Pro1,468.271

MedCodeOverall. Higher is better. No documented setup difference

MiMo-V2.531.895 %

MiMo-V2.5-Pro32.484 %

MedScribeOverall. Higher is better. No documented setup difference

MiMo-V2.572.151 %

MiMo-V2.5-Pro83.73 %

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for MiMo-V2.5 and MiMo-V2.5-Pro
Benchmark and task	MiMo-V2.5	MiMo-V2.5-Pro	Source
Code MigrationOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	14.228 %	Winner21.557 %	Vals AI Source details Data as of: 09/03/2026
CorpFin v2Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	59.907 %	Winner61.422 %	Vals AI Source details Data as of: 08/12/2026
EMBOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	55.092 %	Winner55.228 %	Vals AI Source details Data as of: 09/03/2026
Finance Agent (v2)Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	36.728 %	Winner41.502 %	Vals AI Source details Data as of: 09/03/2026
GPQA DiamondOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	81.566 %	Winner82.576 %	Vals AI Source details Data as of: 09/01/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved Test details No documented difference in the test setupVersion: 1Metric: task resolution rateScoring: Higher is better	1.667 %	Winner2.083 %	Vals AI Source details Data as of: 09/03/2026
Legal Research BenchOverall · All-pass Test details No documented difference in the test setupVersion: 1Metric: all-pass rateScoring: Higher is better	9.135 %	Winner15.865 %	Vals AI Source details Data as of: 09/03/2026
LegalBenchOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner78.886 %	77.129 %	Vals AI Source details Data as of: 09/01/2026
LiveCodeBench v6Overall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	Winner81.512 %	81.354 %	Vals AI Source details Data as of: 09/01/2026
LMArena Text, Style ControlOverall Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 44,540 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 57,416 anonymous pairwise votes.	1,433.798 95% confidence interval: ±4.342The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,468.271 95% confidence interval: ±3.92The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MedCodeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	31.895 %	Winner32.484 %	Vals AI Source details Data as of: 09/03/2026
MedScribeOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	72.151 %	Winner83.73 %	Vals AI Source details Data as of: 09/03/2026
MMLU-ProOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	82.931 %	Winner84.595 %	Vals AI Source details Data as of: 09/01/2026
ProofBenchOverall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is better	16 %	Winner22 %	Vals AI Source details Data as of: 09/01/2026
SWE-bench VerifiedOverall Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is better	71 %	Winner74 %	Vals AI Source details Data as of: 09/01/2026
TaxEval v2Overall Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is better	71.832 %	Winner73.794 %	Vals AI Source details Data as of: 09/01/2026
Terminal-Bench 2.1Overall Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is better	Winner60.674 %	57.303 %	Vals AI Source details Data as of: 09/03/2026
Vals IndexOverall Test details No documented difference in the test setupVersion: 2Metric: weighted index scoreScoring: Higher is better	39.91 %	Winner40.974 %	Vals AI Source details Data as of: 09/03/2026
Vibe Code Bench v1.1Overall Test details No documented difference in the test setupVersion: 1.1Metric: accuracyScoring: Higher is betterSettings: Harness: OpenHands	Winner42.174 %	34.113 %	Vals AI Source details Data as of: 09/03/2026
Code MigrationCLI Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	9.88 %	Winner13.981 %	Vals AI Source details Data as of: 08/09/2026
CorpFin v2Exact Pages Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	59.441 %	Winner60.373 %	Vals AI Source details Data as of: 08/06/2026
Finance Agent (v2)Adjustments Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	31.746 %	Winner34.386 %	Vals AI Source details Data as of: 08/06/2026
GPQA DiamondFew-Shot CoT Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	79.798 %	Winner80.808 %	Vals AI Source details Data as of: 08/09/2026
LegalBenchConclusion Tasks Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	86.386 %	Winner88.031 %	Vals AI Source details Data as of: 08/09/2026
LiveCodeBench v6Easy Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	98.447 %	Winner99.068 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlBusiness, management and finance Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 9,127 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 11,336 anonymous pairwise votes.	1,441.542 95% confidence interval: ±7.42The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,471.792 95% confidence interval: ±6.73The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MMLU-ProBiology Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	90.237 %	Winner92.469 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified<15 min fix Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	80.928 %	Winner82.474 %	Vals AI Source details Data as of: 08/08/2026
TaxEval v2Correctness Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	60.752 %	Winner63.123 %	Vals AI Source details Data as of: 08/09/2026
Terminal-Bench 2.1Easy Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner100 %	75 %	Vals AI Source details Data as of: 08/06/2026
Vals IndexCorpFin v2 Test details No documented difference in the test setupVersion: 1.2Metric: benchmark scoreScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	60.14 %	Winner62.471 %	Vals AI Source details Data as of: 08/09/2026
Code MigrationCOBOL Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	27.273 %	Winner44.286 %	Vals AI Source details Data as of: 08/09/2026
CorpFin v2Max Fitting Context Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	60.14 %	Winner61.422 %	Vals AI Source details Data as of: 08/06/2026
Finance Agent (v2)All-Pass Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	23.785 %	Winner28.848 %	Vals AI Source details Data as of: 08/06/2026
GPQA DiamondZero-Shot CoT Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	83.333 %	Winner84.343 %	Vals AI Source details Data as of: 08/09/2026
LegalBenchInterpretation Tasks Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	74.351 %	Winner79.497 %	Vals AI Source details Data as of: 08/09/2026
LiveCodeBench v6Hard Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner59.143 %	58.571 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlChinese Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 2,140 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 3,151 anonymous pairwise votes.	1,483.45 95% confidence interval: ±13.352The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,511.885 95% confidence interval: ±11.287The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MMLU-ProBusiness Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	89.227 %	Winner89.987 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified>4 hours Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	0 %	Winner33.333 %	Vals AI Source details Data as of: 08/08/2026
TaxEval v2Stepwise Reasoning Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	82.911 %	Winner84.464 %	Vals AI Source details Data as of: 08/09/2026
Terminal-Bench 2.1Hard Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	36.667 %	Winner40 %	Vals AI Source details Data as of: 08/06/2026
Vals IndexFinance Agent v2 Test details No documented difference in the test setupVersion: 1.2Metric: benchmark scoreScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	37.61 %	Winner41.243 %	Vals AI Source details Data as of: 08/09/2026
Code MigrationCode Quality Test details No documented difference in the test setupVersion: 1Metric: mean reviewer code-quality ratingScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	2.788 / 5	Winner3.382 / 5	Vals AI Source details Data as of: 08/09/2026
CorpFin v2Shared Max Context Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	60.14 %	Winner62.471 %	Vals AI Source details Data as of: 08/06/2026
Finance Agent (v2)Comparables Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	31.672 %	Winner40.15 %	Vals AI Source details Data as of: 08/06/2026
LegalBenchIssue Tasks Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner86.337 %	80.173 %	Vals AI Source details Data as of: 08/09/2026
LiveCodeBench v6Medium Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner86.945 %	86.423 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlCoding Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 12,854 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 15,755 anonymous pairwise votes.	1,491.02 95% confidence interval: ±6.676The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,520.349 95% confidence interval: ±6.131The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MMLU-ProChemistry Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	89.841 %	Winner90.459 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified1-4 hours Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Tie33.333 %	Tie33.333 %	Vals AI Source details Data as of: 08/08/2026
Terminal-Bench 2.1Medium Test details No documented difference in the test setupVersion: 2.1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner70.909 %	65.455 %	Vals AI Source details Data as of: 08/06/2026
Vals IndexSWE-bench Test details No documented difference in the test setupVersion: 1.2Metric: benchmark scoreScoring: Higher is better	64.706 %	Winner66.667 %	Vals AI Source details Data as of: 08/09/2026
Finance Agent (v2)Disclosure Analysis Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	47.346 %	Winner53.148 %	Vals AI Source details Data as of: 08/06/2026
LegalBenchRhetoric Tasks Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner77.776 %	64.964 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlCreative writing Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 7,343 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 10,262 anonymous pairwise votes.	1,392.817 95% confidence interval: ±8.004The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,435.028 95% confidence interval: ±7.065The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MMLU-ProComputer Science Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	86.585 %	Winner88.049 %	Vals AI Source details Data as of: 08/09/2026
SWE-bench Verified15 min - 1 hour Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	70.498 %	Winner74.713 %	Vals AI Source details Data as of: 08/08/2026
Vals IndexTerminal-Bench 2.1 Test details No documented difference in the test setupVersion: 1.2Metric: benchmark scoreScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner60.674 %	57.303 %	Vals AI Source details Data as of: 08/09/2026
Finance Agent (v2)Earnings Analysis Test details No documented difference in the test setupVersion: 2Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	45.006 %	Winner54.614 %	Vals AI Source details Data as of: 08/06/2026
LegalBenchRule Tasks Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	69.582 %	Winner72.98 %	Vals AI Source details Data as of: 08/09/2026
LMArena Text, Style ControlEnglish Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena text, style control MiMo-V2.5-Pro: Snapshot: mimo-v2.5-pro, Harness: Arena text, style controlMiMo-V2.5: Style-controlled Arena rating from 20,213 anonymous pairwise votes.MiMo-V2.5-Pro: Style-controlled Arena rating from 24,800 anonymous pairwise votes.	1,452.388 95% confidence interval: ±5.607The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	Winner1,481.475 95% confidence interval: ±5.077The Arena score measures human preference rather than a fixed capability test. The confidence interval and model configuration are part of the result.	LMArena Source details Data as of: 09/01/2026
MMLU-ProEconomics Test details No documented difference in the test setupVersion: 1Metric: accuracyScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	86.73 %	Winner87.915 %	Vals AI Source details Data as of: 08/09/2026
Vals IndexVibe Code Bench Test details No documented difference in the test setupVersion: 1.2Metric: benchmark scoreScoring: Higher is betterSettings: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Winner48.148 %	36.321 %	Vals AI Source details Data as of: 08/09/2026
Key to the marks:better valuebetter, but not conclusivetie

How to read this comparison

Winning one benchmark is not an overall verdict

Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.

The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.

Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.

More matchups

Compare MiMo-V2.5 and MiMo-V2.5-Pro with other leading models using the same data and benchmark logic.

View the complete LLM comparison