Head-to-head comparison

Muse Wonder vs. MiMo-V2.5

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

shared result series: 131
distinct benchmarks: 10
latest retrieval: 2026-10-15

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 Muse Wonder and MiMo-V2.5
Attribute	Muse Wonder	MiMo-V2.5
Provider	Meta	Xiaomi
Released	Oct 12, 2026	Apr 23, 2026
Availability	Preview	Open Source
Model type	Proprietary	Open Weights
Parameters	Not published	310B, 15B active
Architecture	Not published	Mixture of Experts
Context window	1,048,576 tokens	1,000,000 tokens
Knowledge cutoff	Not published	Not published
Technical sources	Model, Context	Model, Context

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 Muse Wonder and MiMo-V2.5
Price type	Muse Wonder	MiMo-V2.5
API input	$4	$0.14
API output	$20	$0.28
Cache read	$0.2	$0.0028
Cache write (5 min.)	Not listed	Not listed
Cache write (1 hr.)	Not listed	Not listed
Price verified	10/15/2026 Source	09/22/2026 Source

Muse WonderMiMo-V2.5

API inputUSD per 1M tokens, base tier

Muse Wonder$4

MiMo-V2.5$0.14

API outputUSD per 1M tokens, base tier

Muse Wonder$20

MiMo-V2.5$0.28

Cache readUSD per 1M tokens, base tier

Muse Wonder$0.2

MiMo-V2.5$0.003

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.

No provider tariff for Muse Wonder is available in the stored OpenRouter snapshot. The official direct price is still shown.

Provider pricing for Muse Wonder and MiMo-V2.5
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
Muse Wonder	Meta	Direct from developerStandard API	Lowest listed price: $4	Lowest listed price: $20	Source10/15/2026
MiMo-V2.5	Xiaomi	Direct from developerStandard API	$0.14	$0.28	Source09/22/2026
	GMICloud	Via OpenRoutergmicloud/fp8	Lowest listed price: $0.119	Lowest listed price: $0.238	Source10/16/2026
	Novita	Via OpenRouternovita/fp8	$0.168	$0.336	Source10/16/2026
	Venice	Via OpenRoutervenice/fp8	$0.4	$2	Source10/16/2026
	Xiaomi	Via OpenRouterxiaomi/fp8	$0.14	$0.28	Source10/16/2026

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.

Not enough directly matched categories for a reliable radar

Only 0 of at least 5 categories have comparable, direct model measurements. The benchmark tables below show the available evidence without filling missing axes.

Muse WonderMiMo-V2.5

Code MigrationOverall. Higher is better. Test setup not documented as identical

Muse Wonder62.552 %

MiMo-V2.514.228 %

EMBOverall. Higher is better. Test setup not documented as identical

Muse Wonder67.084 %

MiMo-V2.555.092 %

Finance Agent (v2)Overall. Higher is better. Test setup not documented as identical

Muse Wonder56.637 %

MiMo-V2.536.728 %

Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. Test setup not documented as identical

Muse Wonder17.083 %

MiMo-V2.51.667 %

Legal Research BenchOverall · All-pass. Higher is better. Test setup not documented as identical

Muse Wonder50.481 %

MiMo-V2.59.135 %

LMArena Code Arena WebDevOverall. Higher is better. Test setup not documented as identical

Muse Wonder1,650.947

MiMo-V2.51,438.179

LMArena Text, Style ControlOverall. Higher is better. Test setup not documented as identical

Muse Wonder1,512.671

MiMo-V2.51,433.804

ProofBenchOverall. Higher is better. Test setup not documented as identical

Muse Wonder84 %

MiMo-V2.516 %

Tax Agent BenchOverall. Higher is better. Test setup not documented as identical

Muse Wonder36.269 %

MiMo-V2.57.772 %

Vibe Code Bench v1.1Overall. Higher is better. Test setup not documented as identical

Muse Wonder83.089 %

MiMo-V2.542.174 %

Code MigrationCLI. Higher is better. Test setup not documented as identical

Muse Wonder60.07 %

MiMo-V2.59.88 %

EMBComps. Higher is better. Test setup not documented as identical

Muse Wonder79.488 %

MiMo-V2.547.165 %

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for Muse Wonder and MiMo-V2.5
Benchmark and task	Muse Wonder	MiMo-V2.5	Source
Code MigrationOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value62.552 %	14.228 %	Vals AI Source details Data as of: 10/12/2026
EMBOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value67.084 %	55.092 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Overall Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value56.637 %	36.728 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value17.083 %	1.667 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchOverall · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value50.481 %	9.135 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevOverall Test details Test setup not documented as identicalNo direct winnerVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Code Arena WebDev MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Code Arena WebDevMuse Wonder: Arena rating from 1,228 published pairwise votes.MiMo-V2.5: Arena rating from 14,953 published pairwise votes.	Higher measured value1,650.947 95% confidence interval: ±18.263The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,438.179 95% confidence interval: ±6.012The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Text, Style ControlOverall Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 1,944 published pairwise votes.MiMo-V2.5: Arena rating from 47,388 published pairwise votes.	Higher measured value1,512.671 95% confidence interval: ±13.236The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,433.804 95% confidence interval: ±4.223The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
ProofBenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value84 %	16 %	Vals AI Source details Data as of: 10/12/2026
Tax Agent BenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value36.269 %	7.772 %	Vals AI Source details Data as of: 10/12/2026
Vibe Code Bench v1.1Overall Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Harness: OpenHands, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Harness: OpenHands, Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value83.089 %	42.174 %	Vals AI Source details Data as of: 10/12/2026
Code MigrationCLI Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value60.07 %	9.88 %	Vals AI Source details Data as of: 10/12/2026
EMBComps Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value79.488 %	47.165 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Adjustments Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value52.67 %	31.746 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkAntitrust Competition · Criteria passed Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: criteria pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value93.841 %	69.565 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchAdministrative / Regulatory · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value63.636 %	10.606 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevWebdev Test details Test setup not documented as identicalNo direct winnerVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Code Arena WebDev MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Code Arena WebDevMuse Wonder: Arena rating from 1,228 published pairwise votes.MiMo-V2.5: Arena rating from 14,953 published pairwise votes.	Higher measured value1,650.947 95% confidence interval: ±18.263The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,438.179 95% confidence interval: ±6.012The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Text, Style ControlBusiness, management and finance Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 338 published pairwise votes.MiMo-V2.5: Arena rating from 9,680 published pairwise votes.	Higher measured value1,500.247 95% confidence interval: ±30.856The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,441.717 95% confidence interval: ±7.222The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchControversy & Precedence Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value20 %	0 %	Vals AI Source details Data as of: 10/12/2026
Code MigrationCOBOL Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value70 %	27.273 %	Vals AI Source details Data as of: 10/12/2026
EMBDataroom Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value79.232 %	74.957 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)All-Pass Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value45.936 %	23.785 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkAntitrust Competition · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value20 %	0 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchAdministrative / Regulatory · Weighted partial credit Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: weighted partial-credit scoreScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value90.571 %	71.929 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevWebdev-html Test details Test setup not documented as identicalNo direct winnerVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Code Arena WebDev MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Code Arena WebDevMuse Wonder: Arena rating from 158 published pairwise votes.MiMo-V2.5: Arena rating from 1,753 published pairwise votes.	Higher measured value1,721.908 95% confidence interval: ±51.47The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,430.016 95% confidence interval: ±14.99The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Text, Style ControlCoding Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 487 published pairwise votes.MiMo-V2.5: Arena rating from 13,665 published pairwise votes.	Higher measured value1,557.956 95% confidence interval: ±26.189The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,490.934 95% confidence interval: ±6.52The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchControversy & Precedence Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value71.362 %	31.723 %	Vals AI Source details Data as of: 10/12/2026
Code MigrationCode Quality Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: mean reviewer code-quality ratingScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value4.061 / 5	2.788 / 5	Vals AI Source details Data as of: 10/12/2026
EMBDCF Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	51.035 %	Higher measured value52.539 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Comparables Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value47.933 %	31.672 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkArbitration International Dispute Resolution · Criteria passed Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: criteria pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value92.473 %	80.645 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchBusiness & Commercial · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value48.837 %	9.302 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevWebdev-react Test details Test setup not documented as identicalNo direct winnerVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Code Arena WebDev MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Code Arena WebDevMuse Wonder: Arena rating from 787 published pairwise votes.MiMo-V2.5: Arena rating from 11,536 published pairwise votes.	Higher measured value1,637.813 95% confidence interval: ±23.199The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,424.743 95% confidence interval: ±6.683The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Text, Style ControlCreative writing Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 427 published pairwise votes.MiMo-V2.5: Arena rating from 7,914 published pairwise votes.	Higher measured value1,485.682 95% confidence interval: ±29.474The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,394.731 95% confidence interval: ±7.76The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchCurrent & Temporal Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value50 %	12.333 %	Vals AI Source details Data as of: 10/12/2026
EMBLBO Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value60.69 %	45.047 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Disclosure Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value67.252 %	47.346 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkArbitration International Dispute Resolution · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	No direct winner0 %	No direct winner0 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchBusiness & Commercial · Weighted partial credit Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: weighted partial-credit scoreScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value86.88 %	75.839 %	Vals AI Source details Data as of: 10/12/2026
LMArena Text, Style ControlEnglish Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 802 published pairwise votes.MiMo-V2.5: Arena rating from 21,390 published pairwise votes.	Higher measured value1,523.395 95% confidence interval: ±20.831The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,452.203 95% confidence interval: ±5.494The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchCurrent & Temporal Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value77.655 %	30.085 %	Vals AI Source details Data as of: 10/12/2026
EMBM&A Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value76.852 %	60.354 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Earnings Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value76.768 %	45.006 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkBanking Finance · Criteria passed Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: criteria pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value94.424 %	89.963 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchCivil Litigation · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value37.037 %	11.111 %	Vals AI Source details Data as of: 10/12/2026
LMArena Text, Style ControlEntertainment, sports and media Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 536 published pairwise votes.MiMo-V2.5: Arena rating from 10,471 published pairwise votes.	Higher measured value1,468.099 95% confidence interval: ±25.703The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,403.039 95% confidence interval: ±7.177The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchFact-Pattern Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value39.535 %	7.209 %	Vals AI Source details Data as of: 10/12/2026
EMBMisc Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value56.936 %	54.045 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Financial Modeling Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value22.515 %	5.271 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkBanking Finance · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value20 %	0 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchCivil Litigation · Weighted partial credit Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: weighted partial-credit scoreScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value89.018 %	71.755 %	Vals AI Source details Data as of: 10/12/2026
LMArena Text, Style ControlExcluding ties Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 1,390 published pairwise votes.MiMo-V2.5: Arena rating from 36,307 published pairwise votes.	Higher measured value1,538.781 95% confidence interval: ±18.91The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,427.653 95% confidence interval: ±5.461The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchFact-Pattern Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value67.563 %	29.256 %	Vals AI Source details Data as of: 10/12/2026
EMBOperating Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value65.354 %	51.541 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)General Qualitative Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value70.779 %	47.992 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkBankruptcy Restructuring · Criteria passed Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: criteria pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value89.965 %	75.433 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchConstitutional / Civil Rights · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value39.474 %	7.895 %	Vals AI Source details Data as of: 10/12/2026
LMArena Text, Style ControlExpert prompts Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 212 published pairwise votes.MiMo-V2.5: Arena rating from 4,800 published pairwise votes.	Higher measured value1,512.579 95% confidence interval: ±38.177The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,470.738 95% confidence interval: ±9.211The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchForms & Filings Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value23.333 %	3.333 %	Vals AI Source details Data as of: 10/12/2026
EMBScratch Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value69.002 %	57.145 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)General Quantitative Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value77.605 %	55.744 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkBankruptcy Restructuring · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	No direct winner0 %	No direct winner0 %	Vals AI Source details Data as of: 10/12/2026
Legal Research BenchConstitutional / Civil Rights · Weighted partial credit Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: weighted partial-credit scoreScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value89.259 %	75.784 %	Vals AI Source details Data as of: 10/12/2026
LMArena Text, Style ControlHard prompts Test details Test setup not documented as identicalNo direct winnerVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterMuse Wonder: Snapshot: muse-wonder-max, Reasoning: max, Harness: Arena Text, style control MiMo-V2.5: Snapshot: mimo-v2.5, Harness: Arena Text, style controlMuse Wonder: Arena rating from 1,218 published pairwise votes.MiMo-V2.5: Arena rating from 31,875 published pairwise votes.	Higher measured value1,533.453 95% confidence interval: ±16.637The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,462.105 95% confidence interval: ±4.961The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
Tax Agent BenchForms & Filings Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterMuse Wonder: Reasoning: max, Temperature: 1, Output limit: 131,072 tokens MiMo-V2.5: Temperature: 1, Top P: 0.95, Output limit: 128,000 tokens	Higher measured value69.85 %	19.069 %	Vals AI Source details Data as of: 10/12/2026
Key to the marks:better valuebetter, but not conclusivetie

Our model tests

Our tests

There is no shared test round for this selection. Results from different rounds are not mixed.

Earlier tests

Pelican on a bicycle

A community classic for free-form SVG drawing.

Muse Wonder

Not tested yet

MiMo-V2.5

The response reached the token limit. No complete recording is available.

Reasoning (requested): high

Current attempt

Run date: Oct 5, 2026
API provider: openrouter
Requested API model: xiaomi/mimo-v2.5
API model reported in response: xiaomi/mimo-v2.5
Reasoning setting: high
Pinned endpoint: Xiaomi | xiaomi/mimo-v2.5-20260422
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Failed
Response time: 136.3 s

Not yet visually reviewed. A successful recording does not confirm correct physics or full compliance with the prompt.

Earlier attempts (3)

Oct 2, 2026API request failedReasoning setting: high
Oct 1, 2026API request failedReasoning setting: high
Sep 7, 2026API request failedReasoning setting: high

Prompt and test conditions

Generate an SVG of a pelican riding a bicycle

Token limit including reasoning: 8,192 tokens

Task origin (Simon Willison)

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

View the complete LLM comparison