BenchLeader

VIALS-AA leaderboard

As of 16 Nov 2026, the top score on VIALS-AA is 67.1% by Claude Fable 5.2 (Anthropic), ahead of Gemini 4 Carbon at 64.0%, across 21 model configurations with a published result (Handshake AI, result dated 9 Nov 2026). Handshake AI's VIALS: 161 questions about life-science images such as gel blots, microscopy and plasmid maps, answered with code and image tools in a sandbox. Run by Artificial Analysis. VIALS-AA is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 16 Nov 2026)

  • Claude Fable 5.2 holds the top VIALS-AA score at 67.1% (Handshake AI, 9 Nov 2026).
  • Gemini 4 Carbon is second on VIALS-AA at 64.0% (Handshake AI, 28 Oct 2026).
  • VIALS-AA: Handshake AI's VIALS: 161 questions about life-science images such as gel blots, microscopy and plasmid maps, answered with code and image tools in a sandbox. Run by Artificial Analysis (Handshake AI).
  • 21 model configurations have a published VIALS-AA result on BenchLeader (BenchLeader, 16 Nov 2026).
  • VIALS-AA does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Handshake AI's VIALS: 161 questions about life-science images such as gel blots, microscopy and plasmid maps, answered with code and image tools in a sandbox. Run by Artificial Analysis.

Published by
Handshake AIdata via Artificial Analysis
Category
Multimodal
Index weight
Reference only
Models
21
Data as of
16 Nov 2026

Source: Artificial Analysis (artificialanalysis.ai). Data taken from the public leaderboard.

21 of 21
#
1NewClaude Fable 5.2max67.1%72.72026-11-09
2Gemini 4 Carbonhigh64.0%68.92026-10-28
3GPT-6 Astramax62.7%67.32026-09-03
4Claude Opus 5.5max61.5%68.42026-09-22
5GPT-6.1 Terramax60.3%64.12026-10-26
6Gemini 4 Argonhigh59.6%65.32026-09-30
7Claude Sonnet 5.5max59.6%64.22026-09-28
8GPT-6.1 Solmax59.0%66.62026-09-29
9Muse Wondermax55.3%60.82026-10-12
10Claude Haiku 5.5max49.1%55.32026-10-07
11Kimi K3.1 Pro45.3%61.82026-10-23
12MiMo-V2.6-Pro43.5%57.52026-09-21
13Inkling 2 Small42.9%57.92026-11-05
14Gemini 3.8 Flashhigh41.0%59.42026-09-02
15DeepSeek V4.1 Flash 1017max40.4%57.72026-10-17
16GLM 5.3 Flash39.8%57.82026-08-26
17Grok 4.7xhigh37.9%57.42026-09-21
18Qwen3.8 Maxmax31.7%58.72026-09-02
19MiniMax M3.1 Flash29.8%54.32026-10-19
20GPT-6 Lunamax27.3%55.12026-09-22
21Mistral Large 4high20.5%48.32026-10-29

Multimodal AI benchmark leaderboard · All LLM benchmarks

Cite as: BenchLeader, “VIALS-AA leaderboard”, data as of 16 Nov 2026, https://www.benchleader.com/benchmarks/aa_vials.

VIALS-AA: questions

What does VIALS-AA measure?
Handshake AI's VIALS: 161 questions about life-science images such as gel blots, microscopy and plasmid maps, answered with code and image tools in a sandbox. Run by Artificial Analysis. Scores are reported in percent of tasks solved; higher is better.
What is the top score on VIALS-AA?
As of 16 Nov 2026, the highest score on VIALS-AA is 67.1% by Claude Fable 5.2 (result dated 9 Nov 2026), ahead of Gemini 4 Carbon at 64.0%. Scores come from Handshake AI and are re-read daily.
Which AI model leads VIALS-AA?
Claude Fable 5.2 leads VIALS-AA with 67.1% as of 16 Nov 2026 (result dated 9 Nov 2026), ahead of Gemini 4 Carbon at 64.0%.
How many models have VIALS-AA results?
21 model configurations have a VIALS-AA result on BenchLeader, all taken from Handshake AI via Artificial Analysis.
Who runs VIALS-AA and how often is it updated?
VIALS-AA is published by Handshake AI. BenchLeader re-reads the published results once a day and records the date each result was published.
Does VIALS-AA count toward the BenchLeader Index?
No. VIALS-AA is shown for reference but left out of the composite index.