activeMathematics
MATH-500
Metric: Score (higher is better)
500-problem subset of the MATH benchmark covering competition mathematics
Models ranked
13
tracked on this benchmark
Score band
98.8 – 75.2
best → lowest tracked
Snapshot trend
+0.04
Oct 9 → Oct 14 · 1 models
Leaderboard
Tracked models ranked by Score (higher is better).
#Model variant and provenanceRelative to leaderScore
198.8
Version: MATH-500 (thinking mode)Harness: Not recordedEvaluator: Not recordedObserved: Oct 13, 2026Confidence: Not recordedSource
298.0
Version: Thinking mode (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
397.8
Version: MATH500, reasoning on (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Oct 8, 2026Confidence: Not recordedSource
496.8
Version: MATH-500 (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Oct 8, 2026Confidence: Not recordedSource
596.1
Version: MATH benchmark (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
696.0
Version: MATH-500Harness: Not recordedEvaluator: Not recordedObserved: Oct 14, 2026Confidence: Not recordedSource
796.0
Version: MATH-500 (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Oct 9, 2026Confidence: Not recordedSource
894.6
Version: From official Microsoft technical report (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
994.3
Version: pass@1 (official DeepSeek model card)Harness: Not recordedEvaluator: Not recordedObserved: Oct 6, 2026Confidence: Not recordedSource
1093.4
Version: MATH benchmark (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
1188.8
Version: MATH-500 (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
1280.6
Version: 4-shot; default TwoTower diffusion decoding at confidence_threshold=0.8, block_size=16, BF16 on 2xH100; evaluator/harness not published on the model cardHarness: Not recordedEvaluator: Not recordedObserved: Jun 25, 2026Confidence: Not recordedSource
1375.2
Version: MATH-500 (accuracy)Harness: Not recordedEvaluator: Not recordedObserved: Jun 7, 2026Confidence: Not recordedSource
How to read this benchmark
This benchmark scores models where higher is better. Use scores for directional filtering and shortlisting, not universal quality ranking; then validate pricing, context window, provider availability, and fit for your workload.
Trust this score when
- There is a fresh timestamped snapshot (or multiple snapshots) for this benchmark.
- The model list covers the same version family you can actually deploy today.
- Top candidates overlap with your required routing and feature requirements.
Be cautious when
- There is only one benchmark snapshot or the dataset appears stale.
- The benchmark metric direction is opposite of your decision objective.
- The score difference between options is narrow and likely within implementation variance.
Related benchmarks
Last reviewed: May 20, 2026