τ²-Bench Telecom ↗95.32%
MiniMax-M2.5 · τ²-Bench Telecom ↗
MiniMax / MODEL EVIDENCE
MiniMax M2.5
25 benchmarks · 26 observations · 25 sources · released 2026-02-12
Scores at a glance
25 benchmarks · large dot: MiniMax M2.5; small dots: closest models where reported
ARC-AGI-1 ↗63.67%
MiniMax-M2.5 · ARC-AGI · Score · % ↗
Terminal-Bench 2.0 ↗42.7%
MiniMax-M2.5 · Terminal Bench · Accuracy mean · % ↗
1 more reported result
42.25% · MiniMax-M2.5 · Terminal Bench · Accuracy mean · % · Epoch AI · Terminal Bench ↗
ARC-AGI-2 ↗4.86%
MiniMax-M2.5 · ARC-AGI-2 · Score · % ↗
CL-bench ↗11.4%
MiniMax-M2.5 · CL-bench · Overall · % ↗
CL-bench Life ↗6.3%
MiniMax-M2.5 · CL-bench Life · Overall · % ↗
MiniMax-M2.5 · ProofBench · Accuracy · % ↗
Text Arena (Coding) ↗1384.13 Elo
MiniMax-M2.5 · WebDev Arena · Elo ↗
Epoch Capabilities Index (ECI) ↗146.59 ECI
MiniMax-M2.5 · General ECI · 26 October 2026 fit ↗
LegalBench ↗79.96 %
MiniMax-M2.5 · LegalBench · Accuracy ↗
CaseLaw v2 ↗53.48 %
MiniMax-M2.5 · CaseLaw · Accuracy ↗
CorpFin v2 ↗59.6 %
MiniMax-M2.5 · CorpFin · Accuracy ↗
MortgageTax & TaxEval v2 ↗68.15 %
MiniMax-M2.5 · TaxEval · Accuracy ↗
MedQA (USMLE) ↗92.53 %
MiniMax-M2.5 · MedQA · Accuracy ↗
AIME 2024–2025 · Vals ↗88.75 %
MiniMax-M2.5 · AIME · Accuracy ↗
GPQA Diamond ↗82.07 %
MiniMax-M2.5 · GPQA Diamond · Accuracy ↗
MMLU-Pro ↗80.09 %
MiniMax-M2.5 · MMLU Pro · Accuracy ↗
Vibe Code Bench ↗14.85 %
MiniMax-M2.5 (OpenHands) · Vibe Code Bench · original task set ↗
SWE-bench Verified ↗74.2 %
MiniMax-M2.5 (Mini-SWE-agent) · SWE-bench Verified · Score ↗
LiveCodeBench ↗79.21 %
MiniMax-M2.5 · LiveCodeBench · Accuracy ↗
Minimax-M2.5 (Mini-SWE-Agent) · SWE Atlas - Refactoring · Score ↗
Minimax M2.5 (Mini-SWE-Agent) · Scale Labs · Score ↗
MiniMax M2.5 · HiL-Bench (Human-in-Loop Benchmark) · Score ↗
ALE-Bench ↗618.17 Performance points
minimax-m2.5 · Average contest performance ↗
Vending-Bench 2 ↗-23.16 USD
MiniMax-M2.5 · End-of-year balance ↗
Closest models by score
Similar positions on shared benchmark leaderboards. Smaller gaps are closer.
GPT OSS 120B →OpenAI
9 of 15 shared benchmark comparisons
GLM 4.7 →Z.ai
9 of 15 shared benchmark comparisons
DeepSeek V3.2 →DeepSeek
9 of 15 shared benchmark comparisons
How similarity is measured
Mean absolute gap in percentile positions on shared benchmark settings; lower is closer. At least 60% overlap is required, with a minimum of three when available. Independent evaluations are used when there are at least three; otherwise published report comparisons are used. Repeated model observations within one setting are averaged before ranking. Writing-style traits and composite indices are excluded. This describes similarity, not an overall quality ranking.
Sources
Every plotted observation links to its evaluation. Reported configurations, benchmark versions and task subsets can differ; similarity describes shared scores, not training lift.
- Arena leaderboard — WebDev evaluation ↗
- 𝜏²-Bench Telecom Benchmark | Artificial Analysis ↗
- Ale Bench — methodology and published results ↗
- ARC-AGI-1 ↗
- ARC-AGI-2 ↗
- CL-bench ↗
- CL-bench Life ↗
- Vending Bench 2 — methodology and published results ↗
- AI benchmarks & capabilities ↗
- HiL-Bench (Human-in-Loop Benchmark) — evaluation methodology and results ↗
- Scale Labs — evaluation methodology and results ↗
- SWE Atlas - Refactoring — evaluation methodology and results ↗
- Terminal-Bench 2.0 leaderboard ↗
- AIME — evaluation methodology and results ↗
- CaseLaw — evaluation methodology and results ↗
- CorpFin — evaluation methodology and results ↗
- GPQA Diamond — evaluation methodology and results ↗
- LiveCodeBench — evaluation methodology and results ↗
- LegalBench — evaluation methodology and results ↗
- MedQA — evaluation methodology and results ↗
- MMLU Pro — evaluation methodology and results ↗
- ProofBench — evaluation methodology and results ↗
- SWE-bench Verified — evaluation methodology and results ↗
- TaxEval — evaluation methodology and results ↗
- Vibe Code Bench — evaluation methodology and results ↗
Names used in source tables
MiniMax M2.5 · MiniMax-M2.5 · MiniMax-M2.5 (Mini-SWE-agent) · MiniMax-M2.5 (OpenHands) · Minimax M2.5 (Mini-SWE-Agent) · Minimax-M2.5 (Mini-SWE-Agent) · minimax-m2.5