Labelbox Complex Reasoning
Models solve reasoning problems in mathematics, programming, and other domains. Labelbox reports overall and domain scores.
Labelbox publishes the results on its leaderboard.
- Made by
- Labelbox
- Released
- July 2025
- Kind
- Language reasoning
- Level
- University
- Scored
- Checked automatically
- Answers
- Short answer
- Tasks
- Kept private
- Measures
- Capability
- Interaction
- Single turn
- Refresh
- One fixed set
- Run by
- A host
- Strongest model tested
- GPT-5 #71 on the capability index
- Last model added
- August 2025
- Scores refreshed
- November 20th, 2026
GPT-5 leads with 89.6% among models released by August 2025 GPT-5 leads with 89.6% among models released by August 2025
Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.
The newest model with a score here came out in August 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 1 GPT-5 89.6%
- 2 o3 86.4%
- 3 Grok 4 80.7%
- 4 o4-mini high 78.5%
- 5 Qwen 3.0 72.4%
- 6
Gemini 2.5 Pro Jun 2025 70.2%
- 7 Claude Opus 4 68.2%
- 8 GPT-4.1 66.4%
- 9 o3-mini 63.4%
- 10 Grok 3 63%
- 11
DeepSeek-R1 May 2025 62.1%
- 12 OpenAI o1 Full 56.3%
- 13
Amazon Nova Premier 50.6%
- 14 Claude 3.7 Thinking 50.5%
Source: Official Labelbox Complex Reasoning leaderboard
- 1 GPT-5 89.6%
- 2 o3 86.4%
- 3 Grok 4 80.7%
- 4 o4-mini high 78.5%
- 5
Gemini 2.5 Pro Jun 2025 70.2%
- 6 Claude Opus 4 68.2%
- 7 GPT-4.1 66.4%
- 8 o3-mini 63.4%
- 9 Grok 3 63%
- 10
DeepSeek-R1 May 2025 62.1%
- 11
Amazon Nova Premier 50.6%
Source: Official Labelbox Complex Reasoning leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 89.6% | Aug 2025 | Official leaderboard | |
| 2 | o3 | 86.4% | Apr 2025 | Official leaderboard | |
| 3 | Grok 4 | 80.7% | Jul 2025 | Official leaderboard | |
| 4 | o4-mini | 78.5% | high | Apr 2025 | Official leaderboard |
| 5 | Qwen 3.0 | 72.4% | Official leaderboard | ||
| 6 | | 70.2% | Jun 2025 | Official leaderboard | |
| 7 | Claude Opus 4 | 68.2% | May 2025 | Official leaderboard | |
| 8 | GPT-4.1 | 66.4% | Apr 2025 | Official leaderboard | |
| 9 | o3-mini | 63.4% | Jan 2025 | Official leaderboard | |
| 10 | Grok 3 | 63% | Apr 2025 | Official leaderboard | |
| 11 | | 62.1% | May 2025 | Official leaderboard | |
| 12 | OpenAI o1 Full | 56.3% | Official leaderboard | ||
| 13 | | 50.6% | Apr 2025 | Official leaderboard | |
| 14 | Claude 3.7 Thinking | 50.5% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-5 | 89.6% | Aug 2025 | Official leaderboard | |
| 2 | o3 | 86.4% | Apr 2025 | Official leaderboard | |
| 3 | Grok 4 | 80.7% | Jul 2025 | Official leaderboard | |
| 4 | o4-mini | 78.5% | high | Apr 2025 | Official leaderboard |
| 5 | | 70.2% | Jun 2025 | Official leaderboard | |
| 6 | Claude Opus 4 | 68.2% | May 2025 | Official leaderboard | |
| 7 | GPT-4.1 | 66.4% | Apr 2025 | Official leaderboard | |
| 8 | o3-mini | 63.4% | Jan 2025 | Official leaderboard | |
| 9 | Grok 3 | 63% | Apr 2025 | Official leaderboard | |
| 10 | | 62.1% | May 2025 | Official leaderboard | |
| 11 | | 50.6% | Apr 2025 | Official leaderboard |
Rankings by part
We read 5 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as math, pure math and applied math. GPT-5 leads every one of them. Arrows show how far a model moved from its overall rank among the models we track.
math
7 models we track have a score here.
- 1 GPT-5 92.2%
- 2 o3 75.3%
- 3 o4-mini ↑1 63.9%
- 4
DeepSeek-R1 May 2025 ↑6 62%
- 5 Claude Opus 4 ↑1 60.8%
- 6 GPT-4.1 ↑1 53.1%
- 7
Amazon Nova Premier ↑4 48.1%
pure math
7 models we track have a score here.
- 1 GPT-5 93%
- 2 o3 91.1%
- 3 o4-mini ↑1 84.6%
- 4 GPT-4.1 ↑3 71.6%
- 5 Claude Opus 4 ↑1 71.1%
- 6
DeepSeek-R1 May 2025 ↑4 65.9%
- 7
Amazon Nova Premier ↑4 41.3%
applied math
7 models we track have a score here.
- 1 GPT-5 86.3%
- 2 o3 77.8%
- 3 o4-mini ↑1 70.7%
- 4 Claude Opus 4 ↑2 61.5%
- 5 GPT-4.1 ↑2 59.8%
- 6
DeepSeek-R1 May 2025 ↑4 58.6%
- 7
Amazon Nova Premier ↑4 52.8%
computer science
7 models we track have a score here.
- 1 GPT-5 88.7%
- 2 o3 82.6%
- 3 Claude Opus 4 ↑3 63.1%
- 4 o4-mini 61.5%
- 5
Amazon Nova Premier ↑6 55.6%
- 6 GPT-4.1 ↑1 52%
- 7
DeepSeek-R1 May 2025 ↑3 49.3%
general reasoning
7 models we track have a score here.
- 1 GPT-5 87.4%
- 2 o3 80.1%
- 3 o4-mini ↑1 78.2%
- 4 Claude Opus 4 ↑2 68.3%
- 5 GPT-4.1 ↑2 66.2%
- 6
DeepSeek-R1 May 2025 ↑4 58.8%
- 7
Amazon Nova Premier ↑4 46.8%
2 models have held number one on the official leaderboard since its release in July 2025
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Labelbox Complex Reasoning leaderboard
O3 on the official leaderboard passed 85% in April 2025
Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Labelbox Complex Reasoning leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
More in Knowledge
- MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects.
Intern-S2-397B 89.8%
- MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy.
Seed1.8 92.3%
- SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy.
Qwen 3.6 Max (Preview) 73.9%
- WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1.
Tencent Hy4 preview 83.9%
- BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly.
Gemma 4 31B IT 74.4%
- HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%
Data audit
All 8 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access The tasks are private.
- Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
- Pass Independent runs Scores from the official leaderboard (14 entries).
- Pass Models scored 11 models we track, plus 3 other entries on its leaderboard.
- Note Newest score The newest came 15 months ago, in August 2025, for GPT-5. That’s expected for a saturated test.
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-5 (#71).
- Note Headroom Saturated. The best is 89.6%, past 85% since April 2025.
Leaderboard import 3 pass
- Pass Importer We read the board on labelbox.com: 11 rows for models we track and 3 other entries at the last read.
- Pass Last read November 20th, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.