BlogLLM Benchmark Scores as Published, Side by Side
LAB-Bench: Results as Published
Every result the catalog holds for LAB-Bench: 14 models from 4 developers, read from 4 pages. Each number is as printed by the company named beside it; nothing is ranked or rescored. Models are listed by developer, newest first.
Area Science and MathMaintainer FutureHouseIntroduced 2024Data version 2026103001, as of 2026-10-30
LAB-Bench is a set of more than 2,400 multiple-choice questions that test practical biology research skills for language models and AI agents. The categories include recalling and reasoning over scientific literature, interpreting figures, using biology databases, and working with DNA and protein sequences. Models are scored on the share of questions they answer correctly and are compared with human expert biology researchers.
LAB-Bench 2
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| GPT-6 AstraOpenAI, released 2026-09-03 | 85.4% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| GPT-5.6 SolOpenAI, released 2026-07-09 | 82.1% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
| GPT-5.6 TerraOpenAI, released 2026-07-09 | 81.2% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Claude Opus 5.5Anthropic, released 2026-09-22 | 73.1% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Claude Fable 5.1Anthropic, released 2026-09-01 | 68.6% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Claude Opus 5Anthropic, released 2026-07-24 | 84.2% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
| Claude Sonnet 5Anthropic, released 2026-06-30 | 80.1% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Gemini 4 ArgonGoogle, released 2026-09-30 | 88.8% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Gemini 3.8 FlashGoogle, released 2026-09-02 | 86.2% | Google, Gemini 3.8 Flash model cardchecked 2026-10-15 | - |
| Gemini 3.7 FlashGoogle, released 2026-08-13 | 82.1% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15; printed on 1 more page | - |
| Gemini 3.6 FlashGoogle, released 2026-07-21 | 76.1% | Google, Gemini 3.7 Flash model cardchecked 2026-10-15 | - |
LAB-Bench (practical MCQ)
| Model | Value | Printed by | Setup as printed |
|---|---|---|---|
| Grok 4.7xAI, released 2026-09-21 | 76.8% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
| Grok 4.6xAI, released 2026-08-12 | 80.7% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
| Grok 4.5xAI, released 2026-07-16 | 71.1% | xAI, Grok 4.7 model cardchecked 2026-10-14 | Effort: high |
Every number here is read from the same catalog file as the benchmark table, data version 2026103001, and is the number the table shows for that model and test. Where one page printed a model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.