A more demanding successor to MMLU emphasizes reasoning across a broad academic subject mix.
161 model results
One row per model and compatible score measure. Mean = equal-weight arithmetic average of imported reports. Distinct configurations reported for the same model in a source table remain separate. Reports can overlap; the range is not a confidence interval. Unbundle to inspect each observation.
87.05 % · MMLU Pro · Accuracy · Cost In / Out: $0.3 / $1.2; Latency: 40.96s [3]
Published observations span 0.44 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#40 of 161 in this displayed comparison; 5.11 pp below the leader.
86.23 % · MMLU Pro · Accuracy · Cost In / Out: $1.75 / $14; Latency: 35.41s [3]
Published observations span 1.17 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#45 of 161 in this displayed comparison; 5.38 pp below the leader.
86.54 % · MMLU Pro · Accuracy · Cost In / Out: $1.25 / $10; Latency: 37.24s [3]
Published observations span 0.53 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#48 of 161 in this displayed comparison; 5.58 pp below the leader.
86.38 % · MMLU Pro · Accuracy · Cost In / Out: $1.25 / $10; Latency: 23.06s [3]
Published observations span 0.66 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#50 of 161 in this displayed comparison; 5.67 pp below the leader.
Publisher-selected recent-model cohort from the public evaluation page. Model labels preserve effort and fallback configurations. Values are a dated snapshot, not a live feed. Consult the publisher’s methodology for the harness, task split and judging; do not merge this cohort with vendor-run results.
Earlier-2026 baseline comparison from the official Qwen model card. Original model labels and score formatting are retained. Language and vision tables use different Qwen reference models. Tools, image inputs, agent scaffolds and task subsets can differ; follow the model card’s evaluation details. Video-MME subtitle settings and HLE tool settings stay in separate rows. Original MCP-Mark results predate MCPMark Verified.
Public source leaderboard captured 26 October 2026. Model names, configurations, native score units and source-reported uncertainty are retained. Version and task split follow this specific Vals methodology page. These are published observations, not evaluations run by RLTasks.
TRACE THE SCORE · Per-model notes use published breakdowns where available.
Reported in recent model launches
These model reports include this benchmark. Their harnesses, task subsets, scoring policies, and resource budgets can differ; follow each report’s evaluation notes before joining the numbers.
A more demanding successor to MMLU emphasizes reasoning across a broad academic subject mix.
How to read the result
Use per-subject scores to detect gaps hidden by the overall average.
Limits and comparison traps
Ten-option questions reduce easy guessing but remain a closed-ended test. Prompt and answer-extraction choices still matter.
Published evaluation
The linked publication documents the task and scoring procedure. Model results retain the source’s evaluation setting, native score units and available uncertainty.
Where this fits in an evaluation suite
Read accuracy together with error and abstention rates. Domain averages can hide language gaps, specialist failures, and poorly calibrated confidence.
Start with a capability hypothesis, then choose a second test with a different failure mode. Keep the checkpoint, reasoning effort, agent tools, sampling policy, and scoring procedure fixed when measuring a training intervention. Record failures and missing outputs in the denominator.
A reproducible comparison
Use the exact implementation, version, task subset, tool access, attempt budget, and judge described by the source.
Publish the exact run configuration, item coverage, score units, and available uncertainty. If a source supplies only an aggregate or a single run, preserve that limit. A new model name or a rounded score alone does not establish a capability improvement.