Benchmarks

36 results across 21 benchmarks

Benchmark results for MiniMax M3
BenchmarkScoreSource
APEX-Agents OriginalEffort: High37.7%mercor.com for MiniMax M3 (high) on APEX-Agents Original (opens in a new tab)Archived copy (opens in a new tab)Pass rateReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on APEX-Agents Original
APEX-Agents Original27.7%huggingface.co for MiniMax M3 on APEX-Agents Original (opens in a new tab)Pass rateReported Model cardMiniMaxSelf-reportedRecord for MiniMax M3 on APEX-Agents Original
AutomationBench 1.0.6Effort: Max4.9%zapier.com for MiniMax M3 (max) on AutomationBench 1.0.6 (opens in a new tab)Archived copy (opens in a new tab)Task pass rateReported Benchmark leaderboardZapierIndependentZapier AutomationBench TeamRecord for MiniMax M3 on AutomationBench 1.0.6
BrowseComp 130-question subset — Mercor web research agentEffort: HighConfiguration: Mercor web research agent, 100 steps, 3 hours53.3%mercor.com for MiniMax M3 (high) on BrowseComp 130-question subset — Mercor web research agent (opens in a new tab)Archived copy (opens in a new tab)BrowseComp AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on BrowseComp 130-question subset — Mercor web research agent
BrowseComp83.5%minimax.io for MiniMax M3 on BrowseComp BrowseComp (opens in a new tab)BrowseComp AccuracyReported Launch postMiniMaxSelf-reportedRecord for MiniMax M3 on BrowseComp BrowseComp
CharXiv Descriptive and reasoning — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent85.9%mercor.com for MiniMax M3 (high) on CharXiv Descriptive and reasoning — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)CharXiv AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on CharXiv Descriptive and reasoning — Mercor single-shot
CL-bench Original20.48%minimax.io for MiniMax M3 on CL-bench Original (opens in a new tab)Task solving rateReported Launch postMiniMaxSelf-reportedRecord for MiniMax M3 on CL-bench Original
DeepSWE 1.1 — Mercor mini-swe-agent, 500 stepsEffort: HighConfiguration: Mercor mini-swe-agent, 500 steps, 2 hours9.1%mercor.com for MiniMax M3 (high) on DeepSWE 1.1 — Mercor mini-swe-agent, 500 steps (opens in a new tab)Archived copy (opens in a new tab)DeepSWE Pass@1Reported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on DeepSWE 1.1 — Mercor mini-swe-agent, 500 steps
Finance Agent Benchmark 2 — October 2026 evaluation48.3%vals.ai for MiniMax M3 on Finance Agent Benchmark 2 — October 2026 evaluation (opens in a new tab)Archived copy (opens in a new tab)Finance Agent AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Finance Agent Benchmark 2 — October 2026 evaluation
GPQA Diamond — Epoch AI harnessEffort: NoneConfiguration: Epoch AI harness81.3%epoch.ai for MiniMax M3 (none) on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Evaluated Benchmark leaderboardEpoch AIIndependentRecord for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness
GPQA Diamond — Epoch AI harnessConfiguration: Epoch AI harness90.9%epoch.ai for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Evaluated Benchmark leaderboardEpoch AIIndependentRecord for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Epoch AI harness
GPQA Diamond — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent90.5%mercor.com for MiniMax M3 (high) on Graduate-Level Google-Proof Q&A Diamond — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Mercor single-shot
GPQA Diamond — Vals AI harnessConfiguration: Vals AI harness: mean of zero-shot and few-shot CoT92.7%vals.ai for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)GPQA Diamond AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Graduate-Level Google-Proof Q&A Diamond — Vals AI harness
HLAB October 2026 evaluation4.2%vals.ai for MiniMax M3 on Harvey's Legal Agent Benchmark October 2026 evaluation (opens in a new tab)Archived copy (opens in a new tab)HLAB Task Pass RateReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Harvey's Legal Agent Benchmark October 2026 evaluation
Humanity's Last Exam Text only — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent37.8%mercor.com for MiniMax M3 (high) on Humanity's Last Exam Text only — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)Humanity's Last Exam AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on Humanity's Last Exam Text only — Mercor single-shot
MedXpertQA MM — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent59.7%mercor.com for MiniMax M3 (high) on MedXpertQA MM — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on MedXpertQA MM — Mercor single-shot
MMLU-Pro Original — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent84.6%mercor.com for MiniMax M3 (high) on MMLU-Pro Original — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on MMLU-Pro Original — Mercor single-shot
MMLU-Pro Original — Vals AI harnessConfiguration: Vals AI harness84.2%vals.ai for MiniMax M3 on MMLU-Pro Original — Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on MMLU-Pro Original — Vals AI harness
MMMU-Pro 1,730 questions — Mercor single-shotEffort: HighConfiguration: Mercor single-shot agent78.9%mercor.com for MiniMax M3 (high) on MMMU-Pro 1,730 questions — Mercor single-shot (opens in a new tab)Archived copy (opens in a new tab)MMMU-Pro accuracySetting: Not statedReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on MMMU-Pro 1,730 questions — Mercor single-shot
MMMU-Pro Vals AI harnessConfiguration: Vals AI harness81.2%vals.ai for MiniMax M3 on MMMU-Pro Vals AI harness (opens in a new tab)Archived copy (opens in a new tab)MMMU-Pro accuracySetting: Standard (4 options)Reported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on MMMU-Pro Vals AI harness
OSWorld Verified — 100 stepsConfiguration: 100-step limit75.2%os-world.github.io for MiniMax M3 on OSWorld Verified — 100 steps (opens in a new tab)Archived copy (opens in a new tab)OSWorld-Verified success rateReported Benchmark leaderboardOSWorldIndependentXLang LabRecord for MiniMax M3 on OSWorld Verified — 100 steps
SWE Atlas Codebase QnA — Mercor mini-swe-agent, 250 stepsEffort: HighConfiguration: Mercor mini-swe-agent, 250 steps, 1 hour37.9%mercor.com for MiniMax M3 (high) on SWE Atlas Codebase QnA — Mercor mini-swe-agent, 250 steps (opens in a new tab)Archived copy (opens in a new tab)Mean pass@1Reported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on SWE Atlas Codebase QnA — Mercor mini-swe-agent, 250 steps
SWE-bench Multilingual — Mercor mini-swe-agent, 1000 stepsEffort: HighConfiguration: Mercor mini-swe-agent, 1000 steps, 3 hours76.0%mercor.com for MiniMax M3 (high) on SWE-bench Multilingual — Mercor mini-swe-agent, 1000 steps (opens in a new tab)Archived copy (opens in a new tab)Resolved TasksReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on SWE-bench Multilingual — Mercor mini-swe-agent, 1000 steps
SWE-bench Verified80.5%huggingface.co for MiniMax M3 on SWE-bench Verified (opens in a new tab)Resolved TasksReported Model cardMiniMaxSelf-reportedRecord for MiniMax M3 on SWE-bench Verified
SWE-bench Verified — Mercor mini-swe-agent, 1000 stepsEffort: HighConfiguration: Mercor mini-swe-agent, 1000 steps, 3 hours75.5%mercor.com for MiniMax M3 (high) on SWE-bench Verified — Mercor mini-swe-agent, 1000 steps (opens in a new tab)Archived copy (opens in a new tab)Resolved TasksReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on SWE-bench Verified — Mercor mini-swe-agent, 1000 steps
SWE-bench Verified — Vals AI mini-swe-agentConfiguration: mini-swe-agent harness75.0%vals.ai for MiniMax M3 on SWE-bench Verified — Vals AI mini-swe-agent (opens in a new tab)Archived copy (opens in a new tab)Resolved TasksReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on SWE-bench Verified — Vals AI mini-swe-agent
SWE-bench Pro Public59.0%minimax.io for MiniMax M3 on SWE-bench Pro Public (opens in a new tab)Public tasks resolvedReported Launch postMiniMaxSelf-reportedRecord for MiniMax M3 on SWE-bench Pro Public
SWE-fficiency Original34.8%minimax.io for MiniMax M3 on SWE-fficiency Original (opens in a new tab)Task success rateReported Launch postMiniMaxSelf-reportedRecord for MiniMax M3 on SWE-fficiency Original
Terminal-Bench 2.0 — Vals AI Terminus 2Configuration: Terminus-2 harness46.1%vals.ai for MiniMax M3 on Terminal-Bench 2.0 — Vals AI Terminus 2 (opens in a new tab)Archived copy (opens in a new tab)Terminal-Bench AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Terminal-Bench 2.0 — Vals AI Terminus 2
Terminal-Bench 2.1 — Mercor Terminus 2Effort: HighConfiguration: Mercor Terminus 2, 1000 steps, 3 hours61.4%mercor.com for MiniMax M3 (high) on Terminal-Bench 2.1 — Mercor Terminus 2 (opens in a new tab)Archived copy (opens in a new tab)Terminal-Bench AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on Terminal-Bench 2.1 — Mercor Terminus 2
Terminal-Bench 2.1 — Terminus-2Configuration: Terminus-2 harness66.0%minimax.io for MiniMax M3 on Terminal-Bench 2.1 — Terminus-2 (opens in a new tab)Terminal-Bench AccuracyReported Launch postMiniMaxSelf-reportedRecord for MiniMax M3 on Terminal-Bench 2.1 — Terminus-2
Terminal-Bench 2.1 — Vals AI Terminus 2Configuration: Terminus-2 harness53.6%vals.ai for MiniMax M3 on Terminal-Bench 2.1 — Vals AI Terminus 2 (opens in a new tab)Archived copy (opens in a new tab)Terminal-Bench AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Terminal-Bench 2.1 — Vals AI Terminus 2
Terminal-Bench 4.0 — Mercor mini-swe-agentEffort: HighConfiguration: Mercor mini-swe-agent, 1000 steps, 8 hours1.0%mercor.com for MiniMax M3 (high) on Terminal-Bench 4.0 — Mercor mini-swe-agent (opens in a new tab)Archived copy (opens in a new tab)Terminal-Bench AccuracyReported Benchmark leaderboardMercorIndependentRecord for MiniMax M3 on Terminal-Bench 4.0 — Mercor mini-swe-agent
Terminal-Bench 4.0 — Vals AI mini-swe-agentConfiguration: mini-swe-agent harness1.0%vals.ai for MiniMax M3 on Terminal-Bench 4.0 — Vals AI mini-swe-agent (opens in a new tab)Archived copy (opens in a new tab)Terminal-Bench AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Terminal-Bench 4.0 — Vals AI mini-swe-agent
Vals Index 2.136.5%vals.ai for MiniMax M3 on Vals Index 2.1 (opens in a new tab)Archived copy (opens in a new tab)Vals Index GDP-weighted AccuracyReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Vals Index 2.1
Vibe Code Bench 1.1 — OpenHands, October 2026 evaluationConfiguration: OpenHands harness47.6%vals.ai for MiniMax M3 on Vibe Code Bench 1.1 — OpenHands, October 2026 evaluation (opens in a new tab)Archived copy (opens in a new tab)Vibe Code Bench ScoreReported Benchmark leaderboardVals AIIndependentRecord for MiniMax M3 on Vibe Code Bench 1.1 — OpenHands, October 2026 evaluation