Academic questions require visual understanding as well as language reasoning, with measures against text-only shortcuts.
124 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.
Published observations span 3.47 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#8 of 114 in this displayed comparison; 3.3 pp below the leader.
Published observations span 5.14 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#9 of 114 in this displayed comparison; 3.33 pp below the leader.
Published observations span 3.47 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#10 of 114 in this displayed comparison; 3.41 pp below the leader.
Published observations span 3.99 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#14 of 114 in this displayed comparison; 4.37 pp below the leader.
Published observations span 5.43 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#16 of 114 in this displayed comparison; 4.51 pp below the leader.
Published observations span 5.14 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#19 of 114 in this displayed comparison; 4.83 pp below the leader.
Published observations span 7.11 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#22 of 114 in this displayed comparison; 5.93 pp below the leader.
Published observations span 7.63 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#23 of 114 in this displayed comparison; 6.3 pp below the leader.
Published observations span 2.3 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#25 of 114 in this displayed comparison; 6.49 pp below the leader.
Published observations span 6.88 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#26 of 114 in this displayed comparison; 6.68 pp below the leader.
Published observations span 6.31 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#29 of 114 in this displayed comparison; 7.04 pp below the leader.
Published observations span 5.78 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#30 of 114 in this displayed comparison; 7.06 pp below the leader.
Published observations span 7.17 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#35 of 114 in this displayed comparison; 7.55 pp below the leader.
Published observations span 6.9 percentage points. Configuration and source differences contribute to this spread; it is not measured run-to-run uncertainty.
#45 of 114 in this displayed comparison; 8.71 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 cohort from the official model card. Asterisks identify competitor scores re-evaluated by Kimi rather than cited from their reports. Coding scores average ten runs; vision averages three runs. Tool-enabled and no-tools results, F1 and exact accuracy, pass@3 and pass^3 remain separate rows. MCPMark here predates the Verified default. APEX uses 452 public tasks.
MMMU-Pro (w/ python) · 2026-10-26
Earlier-2026 baseline cohort from the official model card. Asterisks identify competitor scores re-evaluated by Kimi rather than cited from their reports. Coding scores average ten runs; vision averages three runs. Tool-enabled and no-tools results, F1 and exact accuracy, pass@3 and pass^3 remain separate rows. MCPMark here predates the Verified default. APEX uses 452 public tasks.
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.
MMMU-Pro · MMMU-Pro w/ python · Kimi K3.1 Pro · complete release comparison · 2026-10-27
Published model comparison, 2026. Kimi K3.1 Pro · complete release comparison. Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation. Benchmarks · full embedded comparison table · row 15; source precision and any reported uncertainty retained. These are source-reported observations, not evaluations run by RLTasks.
MMMU-Pro · Kimi K3.1 Pro · complete release comparison · 2026-10-27
Published model comparison, 2026. Kimi K3.1 Pro · complete release comparison. Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation. Benchmarks · full embedded comparison table · row 41; source precision and any reported uncertainty retained. These are source-reported 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.
Academic questions require visual understanding as well as language reasoning, with measures against text-only shortcuts.
How to read the result
Include this when evaluating LLMs with image input, and label the visual setting clearly.
Limits and comparison traps
This is an input capability of a language model. Text-only models and image-input runs need separate comparison cohorts.
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.
Published evaluation
Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation.
Where this fits in an evaluation suite
Ask what the task actually requires: recall, deduction, search, or a mixture. Separate closed-book reasoning from retrieval and tool-assisted problem solving.
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
Exact values from the publisher’s embedded full comparison table. Maximum effort, fallback settings and dated model versions remain in model labels. Blank cells stay missing. Source footnotes specify mixed harnesses and externally quoted comparisons; this is a launch-reported cohort rather than a new independent evaluation.
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.