Skip to main content
BenchLM
Data

MMMU-Pro Extended — Mercor evaluation (MMMU-Pro Extended (Mercor))

We show this table for reference; we do not rank on it.

Answering questions that require both visual and textual understanding. This table shows Mercor-run configurations for reference and is excluded from model rankings.

Evaluation results published by Mercor. Published evaluation aggregates only; task contents and dataset license grants are not included.

Pass@1 on MMMU-Pro Extended (Mercor) — October 15, 2026 capture

We mirror the published pass@1 view for MMMU-Pro Extended (Mercor). GPT-6.1 Sol leads the public snapshot at 61.30%, followed by GPT-6 Astra (60.00%) and GPT-5.6 Sol (Pro) (51.90%). We do not use these results to rank models overall.

1

GPT-6.1 Sol

OpenAI

Mercor extended-set run · max reasoning

Published margin ±7.6 · Original set 85.8% ±1.6 · 408 reported samples

61.30%
2

GPT-6 Astra

OpenAI

Mercor extended-set run · max reasoning

Published margin ±7.5 · Original set 87.5% ±1.5 · 419 reported samples

60.00%
3

GPT-5.6 Sol (Pro)

OpenAI

Mercor extended-set run · max reasoning

Published margin ±7.5 · Original set not reported · 419 reported samples

51.90%

38 configurationsMultimodal & GroundedDated Mercor configurationsDisplay onlyUpdated October 15, 2026 capture

Pass@1 table (38 configurations)

Pass@1 results for MMMU-Pro Extended (Mercor)
RankModel / configurationScoreParameters (B)Open / closed
1GPT-6.1 SolOpenAIMercor extended-set run · max reasoningPublished margin ±7.6 · Original set 85.8% ±1.6 · 408 reported samples
61.30%
Not reportedUnknown
2GPT-6 AstraOpenAIMercor extended-set run · max reasoningPublished margin ±7.5 · Original set 87.5% ±1.5 · 419 reported samples
60.00%
Not reportedUnknown
3GPT-5.6 Sol (Pro)OpenAIMercor extended-set run · max reasoningPublished margin ±7.5 · Original set not reported · 419 reported samples
51.90%
Not reportedUnknown
4GPT-6 SolOpenAIMercor extended-set run · max reasoningPublished margin ±7.6 · Original set 83% ±1.7 · 420 reported samples
51.90%
Not reportedUnknown
5Opus 5.5AnthropicMercor extended-set run · max reasoningPublished margin ±7.6 · Original set 87.4% ±1.5 · 419 reported samples
51.00%
Not reportedUnknown
6Fable 5.1AnthropicMercor extended-set run · max reasoningPublished margin ±7.4 · Original set 86.1% ±1.6 · 420 reported samples
47.60%
Not reportedUnknown
7GPT-5.6 TerraOpenAIMercor extended-set run · max reasoningPublished margin ±7.6 · Original set 80.9% ±1.7 · 420 reported samples
47.40%
Not reportedUnknown
8Sonnet 5.5AnthropicMercor extended-set run · max reasoningPublished margin ±7.8 · Original set 85.8% ±1.6 · 393 reported samples
44.40%
Not reportedUnknown
9GPT-5.5OpenAIMercor extended-set run · xhigh reasoningPublished margin ±7.4 · Original set 81.8% ±1.7 · 420 reported samples
43.80%
Not reportedUnknown
10GPT-5.4OpenAIMercor extended-set run · xhigh reasoningPublished margin ±6.9 · Original set 80.8% ±1.7 · 420 reported samples
39.80%
Not reportedUnknown
11GPT-5.6 LunaOpenAIMercor extended-set run · max reasoningPublished margin ±7 · Original set 79.1% ±1.8 · 416 reported samples
39.20%
Not reportedUnknown
12Fable 5AnthropicMercor extended-set run · max reasoningPublished margin ±7.1 · Original set 72.7% ±2.1 · 420 reported samples
37.60%
Not reportedUnknown
13Gemini 3.7 FlashGoogleMercor extended-set run · high reasoningPublished margin ±6.5 · Original set 85.2% ±1.6 · 420 reported samples
36.40%
Not reportedUnknown
14Gemini 3.8 FlashGoogleMercor extended-set run · high reasoningPublished margin ±6.7 · Original set 85.3% ±1.6 · 420 reported samples
36.40%
Not reportedUnknown
15Kimi K3KimiMercor extended-set run · max reasoningPublished margin ±6.4 · Original set 80.7% ±1.7 · 420 reported samples
36.00%
Not reportedUnknown
16Muse Spark 1.1MetaMercor extended-set run · xhigh reasoningPublished margin ±6.6 · Original set 81.3% ±1.7 · 410 reported samples
35.50%
Not reportedUnknown
17Qwen3.8-MaxAlibabaMercor extended-set run · xhigh reasoningPublished margin ±6.8 · Original set 82.8% ±1.7 · 420 reported samples
34.80%
Not reportedUnknown
18GPT-6 LunaOpenAIMercor extended-set run · max reasoningPublished margin ±6.8 · Original set 76.6% ±1.9 · 420 reported samples
34.80%
Not reportedUnknown
19Opus 5AnthropicMercor extended-set run · high reasoningPublished margin ±7 · Original set not reported · 420 reported samples
34.30%
Not reportedUnknown
20Fable 5AnthropicMercor extended-set run · high reasoningPublished margin ±6.8 · Original set not reported · 420 reported samples
34.30%
Not reportedUnknown
21Muse Spark 1.3MetaMercor extended-set run · xhigh reasoningPublished margin ±6.5 · Original set 81.8% ±1.7 · 420 reported samples
33.80%
Not reportedUnknown
22Haiku 5.5AnthropicMercor extended-set run · max reasoningPublished margin ±6.2 · Original set 79.1% ±1.8 · 420 reported samples
31.90%
Not reportedUnknown
23Opus 4.8AnthropicMercor extended-set run · high reasoningPublished margin ±6.5 · Original set not reported · 420 reported samples
31.40%
Not reportedUnknown
24Gemini 3.1 ProGoogleMercor extended-set run · high reasoningPublished margin ±6.4 · Original set 84.3% ±1.6 · 420 reported samples
31.00%
Not reportedUnknown
25Muse Spark 1.2MetaMercor extended-set run · xhigh reasoningPublished margin ±6.2 · Original set 81.4% ±1.7 · 420 reported samples
30.20%
Not reportedUnknown
26Grok 4.5xAIMercor extended-set run · high reasoningPublished margin ±6.3 · Original set 79% ±1.8 · 420 reported samples
29.80%
Not reportedUnknown
27Gemini 3.6 FlashGoogleMercor extended-set run · high reasoningPublished margin ±6 · Original set 83.7% ±1.7 · 420 reported samples
29.00%
Not reportedUnknown
28Gemini 3.5 FlashGoogleMercor extended-set run · high reasoningPublished margin ±6.3 · Original set 84.7% ±1.6 · 419 reported samples
28.80%
Not reportedUnknown
29Opus 4.7AnthropicMercor extended-set run · high reasoningPublished margin ±6.4 · Original set 78.6% ±1.8 · 420 reported samples
28.30%
Not reportedUnknown
30Grok 4.6xAIMercor extended-set run · xhigh reasoningPublished margin ±6.2 · Original set 80.5% ±1.8 · 420 reported samples
26.00%
Not reportedUnknown
31DeepSeek-V4.1-FlashDeepSeekMercor extended-set run · max reasoningPublished margin ±5.7 · Original set 76.6% ±1.9 · 420 reported samples
25.00%
Not reportedUnknown
32Qwen3.5AlibabaMercor extended-set runPublished margin ±5.4 · Original set not reported · 420 reported samples
24.30%
Not reportedUnknown
33InklingThinking MachinesMercor extended-set run · high reasoningPublished margin ±5.8 · Original set 69.7% ±2 · 420 reported samples
24.00%
Not reportedUnknown
34Kimi K2.7 CodeKimiMercor extended-set run · high reasoningPublished margin ±5.5 · Original set not reported · 420 reported samples
23.10%
Not reportedUnknown
35GLM-5.3-FlashZhipuMercor extended-set run · max reasoningPublished margin ±5.2 · Original set 78.4% ±1.8 · 420 reported samples
22.60%
Not reportedUnknown
36Sonnet 5AnthropicMercor extended-set run · high reasoningPublished margin ±5.4 · Original set not reported · 420 reported samples
20.70%
Not reportedUnknown
37Sonnet 4.6AnthropicMercor extended-set run · high reasoningPublished margin ±5.5 · Original set 74.6% ±2 · 424 reported samples
19.90%
Not reportedUnknown
38MiniMax-M3MiniMaxMercor extended-set run · high reasoningPublished margin ±4.6 · Original set 78.9% ±1.7 · 420 reported samples
18.10%
Not reportedUnknown

How to read this Mercor evaluation

We captured Mercor's MMMU-Pro Extended results on October 15, 2026. The table preserves 38 published configurations, their source identifiers, reasoning settings, and disclosed pass@1 values. These are results from Mercor's evaluation setup; we did not rerun them.

The headline uses the 140 Mercor-built tasks. The paired original-set values refer to the separately listed 1730-task public benchmark. Missing baselines remain unreported. We keep the extension and original set separate from their upstream benchmark-owner and provider-run tables.

The headline is the source page's selected Pass@1 metric. The task set, harness, tools, and grader belong to this Mercor run and may differ from another evaluator's results.

The row labels retain published error margins and sample counts when supplied. Missing margins and counts remain unreported; a sample count is not used to reconstruct a task denominator. Except where the dedicated source defines an interval, the confidence level is unreported. Equal displayed scores and overlapping margins do not establish statistical ties. Capture time does not establish individual evaluation dates.

The general methodology lists programmatic multiple-choice answer extraction and a judge of None. Its run-count label is k = 3. Those notes describe the original public benchmark; extension-specific changes are not established by that reference.

Snapshot

Pass@138 configurationsMercor task extensionDisplay only
Mercor MMMU-Pro Extended resultsMercor APEX benchmark indexMercor methodologyPaired Mercor original-set resultsSource paperUpstream evaluation codeUpstream dataset

The published MMMU-Pro Extended (Mercor) snapshot places GPT-6.1 Sol first at 61.30%. The third row is 9.40 points behind. The broader top-10 range is 21.50 points, so the table still separates the published systems.

38 configurations are shown for MMMU-Pro Extended (Mercor). The benchmark falls in the Multimodal & Grounded category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. MMMU-Pro Extended (Mercor) is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MMMU-Pro Extended (Mercor)

Tasks

140 Mercor tasks; 1730 tasks in the separately reported original set

Format

Pass@1

Difficulty

Source-specific evaluation

The selected headline is Pass@1. Source model identifiers, effort settings, published margins, and reported sample counts remain attached to the source configuration. Public-set results and Mercor task extensions occupy separate tables.

MMMU-Pro Extended evaluation resultsPublic benchmark source

Freshness and provenance

Version

MMMU-Pro Extended; Mercor Pass@1

Refresh cadence

Reviewed dated source capture

Staleness state

Dated Mercor configurations

Question availability

Upstream task access and terms; only aggregate results mirrored

Dated Mercor configurationsDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does MMMU-Pro Extended (Mercor) measure?

Answering questions that require both visual and textual understanding. This table shows Mercor-run configurations for reference and is excluded from model rankings.

Which model leads the published MMMU-Pro Extended (Mercor) snapshot?

GPT-6.1 Sol currently leads the published MMMU-Pro Extended (Mercor) snapshot with 61.30% pass@1. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on MMMU-Pro Extended (Mercor)?

The October 15, 2026 capture snapshot contains 38 source configurations.

Last updated: October 15, 2026 capture · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 5,600+ readers.

One email each week. Unsubscribe anytime.