scBench (original study)
Agents complete 394 tasks in single-cell data analysis. The authors compare their answers with results from the original analyses.
LatchBio distinguishes the original 394 tasks from the current 195. Its current board lists 42 model and harness configurations.
- Made by
- LatchBio
- Released
- February 2026
- Kind
- Genomics and life sciences
- Level
- Expert
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- Public, plus a private test set
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Code sandbox
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- A host
- In views
- Agents, Biology and medicine
- Strongest model tested
- Claude Opus 4.6 #39 on the capability index
- Last model added
- February 2026
- Scores refreshed
- November 21st, 2026
Claude Opus 4.6 leads with 52.8% among models released by February 2026 Claude Opus 4.6 leads with 52.8% among models released by February 2026
Best score per entry from the benchmark's paper, read on November 21st, 2026. A paper’s table doesn’t add newer models. 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 February 2026, 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 Claude Opus 4.6 52.8%
- 2 Claude Opus 4.5 49.9%
- 3 GPT-5.2 45.2%
- 4 Claude Sonnet 4.5 44.2%
- 5 GPT-5.1 37.9%
- 6 Grok 4.1 35.6%
- 7 Grok 4 33.9%
- 8 Gemini 2.5 Pro (snapshot unspecified) 29.2%
Source: The scBench (original study) paper
- 1 Claude Opus 4.6 52.8%
- 2 Claude Opus 4.5 49.9%
- 3 GPT-5.2 45.2%
- 4 Claude Sonnet 4.5 44.2%
- 5 GPT-5.1 37.9%
- 6 Grok 4.1 35.6%
- 7 Grok 4 33.9%
Source: The scBench (original study) paper
| # | Entry | Score ± | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 | 52.8% ±4.5 | Feb 2026 | Official leaderboard | |
| 2 | Claude Opus 4.5 | 49.9% ±4.6 | Nov 2025 | Official leaderboard | |
| 3 | GPT-5.2 | 45.2% ±4.3 | Dec 2025 | Official leaderboard | |
| 4 | Claude Sonnet 4.5 | 44.2% ±4.4 | Sep 2025 | Official leaderboard | |
| 5 | GPT-5.1 | 37.9% ±4.1 | Nov 2025 | Official leaderboard | |
| 6 | Grok 4.1 | 35.6% ±4.1 | Nov 2025 | Official leaderboard | |
| 7 | Grok 4 | 33.9% ±3.8 | Jul 2025 | Official leaderboard | |
| 8 | Gemini 2.5 Pro (snapshot unspecified) | 29.2% ±3.6 | Official leaderboard |
| # | Model | Score ± | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 | 52.8% ±4.5 | Feb 2026 | Official leaderboard | |
| 2 | Claude Opus 4.5 | 49.9% ±4.6 | Nov 2025 | Official leaderboard | |
| 3 | GPT-5.2 | 45.2% ±4.3 | Dec 2025 | Official leaderboard | |
| 4 | Claude Sonnet 4.5 | 44.2% ±4.4 | Sep 2025 | Official leaderboard | |
| 5 | GPT-5.1 | 37.9% ±4.1 | Nov 2025 | Official leaderboard | |
| 6 | Grok 4.1 | 35.6% ±4.1 | Nov 2025 | Official leaderboard | |
| 7 | Grok 4 | 33.9% ±3.8 | Jul 2025 | Official leaderboard |
2 models have held number one in the benchmark’s paper since its release in February 2026
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The scBench (original study) paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official scBench (original study) 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.
Versions
More in Biology and chemistry
- GeneBench Pro In use Agents analyze messy simulated genomics and biology datasets to estimate a decision-relevant quantity across 129 problems. Scored automatically against known answers. GPT-5.6 Sol 31.5%
- MedChemBench (OpenAI) In use Models work through 99 medicinal chemistry problems drawn from recent papers and patents. Each task mixes prose with structures, assay tables and plots, then asks for the program’s next decision. GPT-6 Astra 83.8%
- SUPERChem In use Models solve 500 expert-written chemistry reasoning problems, offered in multimodal and text-only forms. Expert solution paths also allow a reasoning path score.
Tencent Hy4 preview 66.4%
- LifeSciBench In use Models complete 750 expert-written life science research tasks, often with attached figures, tables, or sequence files. Graded against expert rubrics as score and pass rate. GPT-Rosalind-6 61.3%
- ProtocolQA Open-Ended In use Models correct errors in published lab protocols using open-ended questions adapted from LAB-Bench ProtocolQA. Scored as pass@1. GPT-6 Cyber 47.2%
- MolecularIQ In use Models answer questions about molecular graphs, such as counting, indexing and constrained generation, and a symbolic checker verifies each answer. gpt-oss-120b 47.5%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access Public tasks, plus a private test set.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the benchmark’s paper (8 entries).
- Pass Models scored 7 models we track, plus 1 other entry on its leaderboard.
- Note Newest score The newest came 9 months ago, in February 2026, for Claude Opus 4.6. That’s expected for a test replaced by scBench (LatchBio SingleCellBench).
- Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Claude Opus 4.6 (#39).
- Pass Headroom The best is 52.8%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on arxiv.org: 7 rows for models we track and 1 other entry at the last read.
- Pass Last read November 21st, 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.