Side by side

See the difference.

Scores, capabilities and costs for the models you’re considering.

Scroll to compare models →

Comparison of MiniMax-M3, MiniMax: MiniMax M2.5, remove_background_guided
3 modelsMiniMaxAIMiniMax-M3Remove ×minimaxMiniMax: MiniMax M2.5Remove ×Briaremove_background_guidedRemove ×
What’s differentMiniMax-M3 is a native multimodal model with 1M context. It has ~428B parameters and ~23B activated parameters.MiniMax-M2.5 is a SOTA large language model designed for real-world productivity. Trained in a diverse range of complex real-world digital working environments, M2.5 builds upon the coding expertise of M2.1...Prompt-guided background removal by Bria: tell it what to keep or what to remove and get a clean cutout with true transparency, at the original resolution with the original pixels untouched. Trained…
DateJun 1, 2026ReleasedFeb 12, 2026ListedOct 14, 2026Listed
Listed rate$0.28 in / $1.10 out/ 1M tokensdeepinfra · via Hugging Face · all routes ↗$0.27 in / $0.95 out/ 1M tokensVenice · via OpenRouter · all routes ↗veniceSee provider quotes
Provider endpoints7 endpoints →12 endpoints →0 endpoints →
AA Intelligence ↗4.3.229.22 ↗Effort not specifiedIndependent resultNot reportedNot reported
AA-Briefcase ↗1.11,090.37 ↗Effort not specifiedIndependent resultNot reportedNot reported
GDPval-AA ↗2.11,242.47 ↗Effort not specifiedIndependent resultNot reportedNot reported
AutomationBench-AA ↗AA Index 4.3.2 collection21.25% ↗Effort not specifiedIndependent resultNot reportedNot reported
Terminal-Bench 4.0 · AA ↗2.02% ↗Effort not specifiedHarness unreportedIndependent resultNot reportedNot reported
SciCode ↗AA Index 4.3.2 collection47.11% ↗AA · Under reviewEffort not specifiedIndependent resultNot reportedNot reported
Humanity’s Last Exam ↗AA Index 4.3.2 collection38.97% ↗Effort not specifiedIndependent resultNot reportedNot reported
GDP.pdf ↗AA Index 4.3.2 collection9.8% ↗Effort not specifiedIndependent resultNot reportedNot reported
CritPt ↗AA Index 4.3.2 collection3.71% ↗AA · Under reviewEffort not specifiedIndependent resultNot reportedNot reported
AA-Omniscience ↗AA Index 4.3.2 collection1.35 ↗Effort not specifiedIndependent resultNot reportedNot reported
AA-LCR ↗1.183% ↗Effort not specifiedIndependent resultNot reportedNot reported
Context window1M204.8K—
Max output—32.8K—
Parameters———
Tool callingSupportedSupportedNot reported
Structured outputSupportedSupportedNot reported
Image inputSupportedNot reportedNot reported
ReasoningSupportedSupportedNot reported
Prompt cachingNot reportedSupportedNot reported
Open weightsSupportedSupportedNot reported

Add another model

Search all models →