# APEX-1 — Mercor evaluation (APEX-1 (Mercor))

> Single-turn professional deliverables graded against expert-written criteria. This table shows Mercor-run configurations for reference and is excluded from model rankings.

Canonical page: https://benchlm.ai/benchmarks/mercorapex1

- Category: [external](/external)
- Last updated: October 15, 2026 capture

Evaluation results published by [Mercor](https://www.mercor.com/apex/apex-v1-leaderboard/). Published evaluation aggregates only; task contents and dataset license grants are not included.

We captured Mercor's APEX-1 results on October 15, 2026. The table preserves 70 published configurations, their source identifiers, reasoning settings, and disclosed mean score values. These are results from Mercor's evaluation setup; we did not rerun them.

This dated source table is display only and does not enter overall or category rankings. Source configurations retain their external identity, including Pro descriptors and different effort settings.

Mean score is rubric-criterion coverage, not the fraction of tasks fully completed. The leaderboard description lists 400 tasks and eight runs, while the general methodology lists 300 tasks and four runs. We preserve the published aggregates and sample counts without choosing a common denominator or combining these results with another APEX edition.

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 expert-authored rubrics, scored criterion by criterion and a judge of LLM judge over expert rubrics. Its run-count label is k = 4. The linked result page takes precedence where its setup differs, as noted above.

- [Mercor APEX-1 results](https://www.mercor.com/apex/apex-v1-leaderboard/)
- [Mercor APEX benchmark index](https://www.mercor.com/apex/)
- [Mercor methodology](https://www.mercor.com/apex/methodology/)
- [Source paper](https://arxiv.org/abs/2509.25721)
- [Upstream evaluation code](https://github.com/Mercor-Intelligence/apex-evals/tree/main/apex-evals-v1-extended)
- [Upstream dataset](https://huggingface.co/datasets/Mercor/APEX-v1)
- [Mercor evaluation notes](https://www.mercor.com/blog/expanding-the-ai-productivity-index-apex/)

## About APEX-1 (Mercor)

- Tasks: Task and run counts differ across source pages
- Format: Mean Score
- Difficulty: Source-specific evaluation
- Paper: [APEX-1 evaluation results](https://www.mercor.com/apex/apex-v1-leaderboard/)

The selected headline is Mean Score. 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.

APEX-1 (Mercor) is currently displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.

## Leaderboard (70 configurations)

| Rank | Model | Configuration | Creator | Score |
|------|-------|---------------|---------|-------|
| 1 | [Muse Spark 1.2](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.3 · 1,593 reported samples | Meta | 75.60% |
| 2 | [Muse Spark 1.3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.3 · 1,599 reported samples | Meta | 74.00% |
| 3 | [Fable 5.1](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.7 · 1,577 reported samples | Anthropic | 73.80% |
| 4 | [Muse Spark 1.1](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.2 · 1,591 reported samples | Meta | 73.80% |
| 5 | [GPT-5.6 Sol (Pro)](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.4 · 1,600 reported samples | OpenAI | 73.10% |
| 6 | [Muse Spark 1.3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.3 · 1,600 reported samples | Meta | 73.00% |
| 7 | [GPT-5.6 Terra](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.4 · 1,598 reported samples | OpenAI | 69.50% |
| 8 | [GPT-6 Astra](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.6 · 1,580 reported samples | OpenAI | 69.30% |
| 9 | [GPT-6.1 Sol](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,436 reported samples | OpenAI | 69.30% |
| 10 | [Gemini 3.8 Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,599 reported samples | Google | 68.80% |
| 11 | [Opus 5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.9 · 1,595 reported samples | Anthropic | 68.60% |
| 12 | [Kimi K3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.3 · 1,600 reported samples | Kimi | 68.40% |
| 13 | [GPT-5.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.4 · 1,600 reported samples | OpenAI | 68.40% |
| 14 | [Grok 4.6](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.5 · 1,600 reported samples | xAI | 67.70% |
| 15 | [GPT 5.4](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.4 · 3,142 reported samples | OpenAI | 67.20% |
| 16 | [Gemini 3.7 Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.6 · 1,600 reported samples | Google | 67.00% |
| 17 | [GPT-5.6 Luna](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.5 · 1,586 reported samples | OpenAI | 66.60% |
| 18 | [Fable 5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.9 · 1,600 reported samples | Anthropic | 66.00% |
| 19 | [Gemini 3.6 Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,600 reported samples | Google | 66.00% |
| 20 | [DeepSeek-V4-Pro-0813](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.5 · 1,591 reported samples | DeepSeek | 66.00% |
| 21 | [Opus 5.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±3.2 · 1,600 reported samples | Anthropic | 65.80% |
| 22 | [Opus 4.6](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,600 reported samples | Anthropic | 65.70% |
| 23 | [GPT-5.4](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.4 · 1,600 reported samples | OpenAI | 65.60% |
| 24 | [DeepSeek-V4.1-Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,600 reported samples | DeepSeek | 65.50% |
| 25 | [Grok 4.7](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.6 · 1,591 reported samples | xAI | 65.40% |
| 26 | [Opus 4.6](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,599 reported samples | Anthropic | 65.30% |
| 27 | [GPT-5.2-Codex](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.6 · 1,588 reported samples | OpenAI | 65.30% |
| 28 | [Qwen3.8-Max](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.5 · 1,559 reported samples | Alibaba | 65.30% |
| 29 | [GPT-5.2](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,599 reported samples | OpenAI | 64.70% |
| 30 | [Haiku 5.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,600 reported samples | Anthropic | 64.70% |
| 31 | [GPT-5.3-Codex](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.4 · 1,599 reported samples | OpenAI | 64.60% |
| 32 | [GPT-5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 3,182 reported samples | OpenAI | 64.50% |
| 33 | [DeepSeek-V4-Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,600 reported samples | DeepSeek | 63.80% |
| 34 | [GPT-6 Sol](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.6 · 1,588 reported samples | OpenAI | 63.60% |
| 35 | [GLM-5.3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.9 · 1,593 reported samples | Zhipu | 62.90% |
| 36 | [GPT-5-Codex](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.6 · 1,593 reported samples | OpenAI | 62.80% |
| 37 | [Opus 4.8](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±3 · 1,600 reported samples | Anthropic | 62.70% |
| 38 | [Qwen 3.8 27B](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · xhigh reasoning · Published margin ±2.7 · 1,580 reported samples | Alibaba | 62.60% |
| 39 | [Opus 4.7](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.8 · 1,600 reported samples | Anthropic | 62.20% |
| 40 | [Gemini 3.5 Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.8 · 1,600 reported samples | Google | 62.20% |
| 41 | [GPT-6 Luna](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.7 · 1,600 reported samples | OpenAI | 62.00% |
| 42 | [GPT-5.1-Codex](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,591 reported samples | OpenAI | 61.90% |
| 43 | [Gemini 3.1 Pro](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,599 reported samples | Google | 61.80% |
| 44 | [Grok 4.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,600 reported samples | xAI | 61.80% |
| 45 | [GLM-5.2](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.5 · 1,600 reported samples | Zhipu | 61.50% |
| 46 | [Gemini 3 Pro](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.4 · 3,200 reported samples | Google | 60.90% |
| 47 | [Gemini 3 Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.3 · 1,600 reported samples | Google | 60.90% |
| 48 | [o3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,598 reported samples | OpenAI | 60.50% |
| 49 | [Kimi K2.7 Code](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,596 reported samples | Kimi | 60.10% |
| 50 | [MiniMax-M3](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,600 reported samples | MiniMax | 59.80% |
| 51 | [DeepSeek-V4-Pro](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±2.7 · 1,594 reported samples | DeepSeek | 59.70% |
| 52 | [GLM-5.3-Flash](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±3.1 · 1,600 reported samples | Zhipu | 58.70% |
| 53 | [Inkling](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,596 reported samples | Thinking Machines | 58.10% |
| 54 | [Grok 4.1](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · fast reasoning · Published margin ±2.6 · 1,600 reported samples | xAI | 57.50% |
| 55 | [Opus 4.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.8 · 3,199 reported samples | Anthropic | 57.30% |
| 56 | [Grok 4](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.7 · 1,592 reported samples | xAI | 57.10% |
| 57 | [Nemotron 3 Ultra](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.6 · 1,600 reported samples | NVIDIA | 56.10% |
| 58 | [Sonnet 4.6](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±3.3 · 1,600 reported samples | Anthropic | 55.10% |
| 59 | [Gemini 3.5 Flash Lite](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,600 reported samples | Google | 54.70% |
| 60 | [Sonnet 5.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±3.7 · 1,598 reported samples | Anthropic | 53.90% |
| 61 | [Kimi K2 (Thinking)](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.5 · 1,517 reported samples | Kimi | 52.40% |
| 62 | [GLM-5.1](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.5 · 1,576 reported samples | Zhipu | 52.40% |
| 63 | [GLM-4.7](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.7 · 1,439 reported samples | Zhipu | 51.70% |
| 64 | [DeepSeek-V3.2](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.7 · 1,599 reported samples | DeepSeek | 51.50% |
| 65 | [Qwen3.5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.5 · 1,600 reported samples | Alibaba | 50.30% |
| 66 | [Sonnet 5](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · max reasoning · Published margin ±3.4 · 1,599 reported samples | Anthropic | 49.60% |
| 67 | [GLM-5 (Thinking)](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.9 · 1,599 reported samples | Zhipu | 49.00% |
| 68 | [Gemma 4 31B](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · Published margin ±2.7 · 1,574 reported samples | Google | 48.10% |
| 69 | [GPT-OSS-120B](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,600 reported samples | OpenAI | 46.40% |
| 70 | [MiniMax-M2.7](https://www.mercor.com/apex/apex-v1-leaderboard/) | Single-shot agent · high reasoning · Published margin ±2.7 · 1,600 reported samples | MiniMax | 45.80% |

## FAQ

### What does APEX-1 (Mercor) measure?

Single-turn professional deliverables graded against expert-written criteria. This table shows Mercor-run configurations for reference and is excluded from model rankings.

### Which model leads the published APEX-1 (Mercor) snapshot?

Muse Spark 1.2 currently leads the published APEX-1 (Mercor) snapshot with a score of 75.60%.

### How many models are evaluated on APEX-1 (Mercor)?

The October 15, 2026 capture contains 70 source configurations.

### Does APEX-1 (Mercor) affect BenchLM's overall score?

Not directly. APEX-1 (Mercor) is still displayed on BenchLM for reference, but it is excluded from the weighted scoring formula.