Benchmark Almanac

TxBench-Antibody-Development

Gemini 4 Carbon 61.1% leads since October 2026
Not in model cards

Agents use evidence to make manufacturing, quality, and clinical development decisions about candidate therapeutic antibodies.

LatchBio lists the evaluation in its V1 inventory.

Made by
LatchBio
Kind
Research science
Level
Expert
Scored
Checked automatically
Answers
Agent task
Measures
Capability
Language
English
In and out
Text → Text, Actions
Interaction
Agentic
Environment
Terminal
Run by
A host
Strongest model tested
Gemini 4 Carbon #3 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
23
Models scored on its own leaderboard

Gemini 4 Carbon leads with 61.1%

Best score per model from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best.

  1. 1 Gemini 4 Carbon 61.1%
  2. 2 Gemini 4 Hydrogen 60.6%
  3. 3 Grok 4.7 39.4%
  4. 4 Claude Opus 5 LSVP 37.2%
  5. 5 Claude Sonnet 5.5 36.7%
  6. 6 GPT-6 Astra 32.8%
  7. 7 Claude Opus 5.5 LSVP 32.8%
  8. 8 Gemini 4 Argon 31.1%
  9. 9 Grok 4.6 28.3%
  10. 10 Claude Fable 5.1 27.8%
  11. 11 GPT-6.1 Sol 27.8%
  12. 12 GPT-6.1 Terra 22.8%
  13. 13 GPT-5.6 Sol 18.9%
  14. 14 GPT-6 Luna 18.9%
  15. 15 Gemini 3.8 Flash 15%
  16. 16 DeepSeek V4.1 Flash 14.4%
  17. 17 Claude Sonnet 5 13.3%
  18. 18 GLM-5.3 10%
  19. 19 Kimi K3 6.11%
  20. 20 Mistral Large 4 6.11%
  21. 21 Nemotron 3 Ultra 5.56%
  22. 22 GPT-5.6 Luna 5.00%
  23. 23 Nemotron 3.5 Lightning 30B-A3B 4.44%

Source: Official TxBench-Antibody-Development leaderboard

#ModelScoreSettingDateSource
1 Gemini 4 Carbon 61.1% Pi Oct 2026 Official leaderboard
2 Gemini 4 Hydrogen 60.6% Pi Nov 2026 Official leaderboard
3 Grok 4.7 39.4% Grok Build Sep 2026 Official leaderboard
4 Claude Opus 5 37.2% LSVP, Claude Code Jul 2026 Official leaderboard
5 Claude Sonnet 5.5 36.7% Pi Sep 2026 Official leaderboard
6 GPT-6 Astra 32.8% Pi Sep 2026 Official leaderboard
7 Claude Opus 5.5 32.8% LSVP, Pi Sep 2026 Official leaderboard
8 Gemini 4 Argon 31.1% Pi Sep 2026 Official leaderboard
9 Grok 4.6 28.3% Grok Build Aug 2026 Official leaderboard
10 Claude Fable 5.1 27.8% Pi Sep 2026 Official leaderboard
11 GPT-6.1 Sol 27.8% Codex Sep 2026 Official leaderboard
12 GPT-6.1 Terra 22.8% Codex Oct 2026 Official leaderboard
13 GPT-5.6 Sol 18.9% Pi Jul 2026 Official leaderboard
14 GPT-6 Luna 18.9% Codex Sep 2026 Official leaderboard
15 Gemini 3.8 Flash 15% Pi Sep 2026 Official leaderboard
16 DeepSeek V4.1 Flash 14.4% Pi Sep 2026 Official leaderboard
17 Claude Sonnet 5 13.3% Pi Jun 2026 Official leaderboard
18 GLM-5.3 10% Pi Aug 2026 Official leaderboard
19 Kimi K3 6.11% Pi Jul 2026 Official leaderboard
20 Mistral Large 4 6.11% Pi Oct 2026 Official leaderboard
21 Nemotron 3 Ultra 5.56% Pi Jun 2026 Official leaderboard
22 GPT-5.6 Luna 5.00% Pi Jul 2026 Official leaderboard
23 Nemotron 3.5 Lightning 30B-A3B 4.44% Pi Aug 2026 Official leaderboard

6 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

NVIDIA Anthropic OpenAI xAI Google DeepMind
Own leaderboard 6 leaders

Source: Official TxBench-Antibody-Development leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Gemini 4 Carbon Grok 4.7 GPT-5.6 Sol Claude Sonnet 5 Nemotron 3 Ultra 0% 20% 40% 60% 80% 100% 2026 2027 Near the ceiling (85%) Gemini 4 Carbon Grok 4.7

Source: Official TxBench-Antibody-Development 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 Science and reasoning

All 203 in Science and reasoning
  • GPQA Diamond In use 198 graduate-level 'Google-proof' multiple-choice questions in biology, chemistry and physics. Gemini 4 Carbon 97%
  • Terminal-Bench-Science In use 70 scientific research-workflow tasks across five science areas, run in Harbor containers. Hidden tests give the resolution rate. GPT-6 Astra 68.1%
  • BioMysteryBench In use Bioinformatics problems on anonymized real datasets that require real analysis with tools. Rubric-graded, with human-solvable and human-difficult splits. Claude Fable 5.2 83.7%
  • FrontierScience In use Models solve physics, chemistry, and biology problems in an Olympiad track graded by short answer and a Research track graded by 10-point rubrics. Intern-S2-397B 87%
  • HiPhO In use Models solve theory problems from 2024 and 2025 physics olympiad exams, some with diagrams. Official marking schemes grade answers and steps, mapped to medal levels. Gemini 3 Pro 78.3%
  • LABBench2 In use 11 biology research sub-tasks (literature, figures, tables, patents, trials and more) with tool access. Macro-averaged accuracy. GPT-5.2 60.9%

Data audit

10 of 11 checks pass, 1 warning. Look first at the Task access check. We don’t know whether the tasks are public. Every benchmark

Catalog record 1 warning 1 pass
  • Warning Task access We don’t know whether the tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (23 entries).
  • Pass Models scored 23 models we track.
  • Pass Newest score The newest is from November 2026, for Gemini 4 Hydrogen.
  • Pass Frontier models 8 of the top 10 have a score, led by Gemini 4 Carbon (#3).
  • Pass Headroom The best is 61.1%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on benchmarks.bio: 36 rows for models we track 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.

How we check