BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

Public Benefits Bench leaderboard

As of 6 Nov 2026, the top score on Public Benefits Bench is 76.9% by Claude Opus 5 (Anthropic), ahead of Claude Fable 5.1 at 74.9%, across 50 model configurations with a published result (Vals AI, result dated 12 Oct 2026). Questions a benefits applicant might ask about SNAP eligibility and process, graded against policy. Public Benefits Bench is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5 holds the top Public Benefits Bench v1.1 score at 76.9% (Vals AI, 12 Oct 2026).
  • Claude Fable 5.1 is second on Public Benefits Bench v1.1 at 74.9% (Vals AI, 12 Oct 2026).
  • Public Benefits Bench v1.1: Navigating SNAP benefits questions for applicants. Run by Vals AI (Vals AI).
  • 50 model configurations have a published Public Benefits Bench v1.1 result on BenchLeader (BenchLeader, 6 Nov 2026).
  • Public Benefits Bench v1.1 does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Navigating SNAP benefits questions for applicants. Run by Vals AI.

Published by
Vals AI
Category
Knowledge
Index weight
Reference only
Models
50
Data as of
6 Nov 2026

Vals AI (vals.ai).

What the test looks like

Questions a benefits applicant might ask about SNAP eligibility and process, graded against policy.

How it is scored

Accuracy, run by Vals AI.

What to keep in mind

US-specific and policy changes over time.

Top 15
  1. 1Claude Opus 576.9%
  2. 2Claude Fable 5.174.9%
  3. 3Muse Wonder Full (max)73.8%
  4. 4Claude Opus 5.570.6%
  5. 5Claude Fable 570.4%
  6. 6Muse Wonder (max)70.2%
  7. 7Gemini 4 Argon (high)69.8%
  8. 8MiMo-V2.6-Pro68.9%
  9. 9Hy468.6%
  10. 10GLM 5.3 (max)68.5%
  11. 11Muse Spark 1.2 (xhigh)68.5%
  12. 12Kimi K3 (max)68.2%
  13. 13Claude Opus 4.868.1%
  14. 14MiMo-V2.6-Flash67.6%
  15. 15Claude Sonnet 5.567.2%
50 of 50
#
1Claude Opus 576.9%65.2$3.952026-10-12
2Claude Fable 5.174.9%66.4$7.482026-10-12
3Muse Wonder Fullmax73.8%62.9$4.642026-10-12
4Claude Opus 5.570.6%63.4$5.952026-10-12
5Claude Fable 570.4%66.5$4.592026-10-12
6Muse Wondermax70.2%62.3$2.462026-10-12
7Gemini 4 Argonhigh69.8%66.0$2.792026-10-12
8MiMo-V2.6-Pro68.9%57.3$0.0822026-10-12
9Hy468.6%55.7$0.2532026-10-12
10GLM 5.3max68.5%59.0$0.9472026-10-12
11Muse Spark 1.2xhigh68.5%61.7$0.5882026-10-12
12Kimi K3max68.2%62.1$0.9132026-10-12
13Claude Opus 4.868.1%60.0$1.892026-10-12
14MiMo-V2.6-Flash67.6%55.0$0.0322026-10-12
15Claude Sonnet 5.567.2%–$4.252026-10-12
16Qwen3.8 Maxmax67.1%58.8$0.9912026-10-12
17Grok 4.6high66.8%60.5$0.9162026-10-12
18GPT-5.6 Solmax66.5%65.4$8.852026-10-12
19Ember-166.2%–$1.022026-10-12
20Claude Sonnet 566.0%55.9$1.292026-10-12
21Grok 4.7xhigh65.6%58.3$2.252026-10-12
22Gemini 3.8 Flashhigh65.3%62.0$1.012026-10-12
23DeepSeek V4.1 Flashmax64.3%59.0$0.0732026-10-12
24MiniMax M364.1%56.4$0.2542026-10-12
25Claude Haiku 5.563.2%–$0.5252026-10-12
26DeepSeek V4 Promax62.9%57.0$0.4602026-10-12
27Claude Sonnet 4.662.5%58.2$1.112026-10-12
28GPT-5.6 Terramax62.4%60.7$1.202026-10-12
29GLM 5.161.8%55.7$0.3412026-10-12
30Muse Spark 1.1xhigh61.3%59.5$0.2782026-10-12
31GPT-5.6 Lunamax61.2%56.3$0.2702026-10-12
32GPT-5.5xhigh60.9%64.3$3.972026-10-12
33Gemini 3.5 Flashhigh59.5%62.5$1.112026-10-12
34Inkling Small59.4%52.1$0.0962026-10-12
35GPT-6.1 Solmax59.3%67.3$0.6912026-10-12
36Inkling58.5%47.5$0.3682026-10-12
37Step 5 Preview58.1%54.9$0.5042026-10-12
38GPT-6 Lunamax57.6%56.7$0.2242026-10-12
39GPT-6 Solmax56.6%63.3$2.492026-10-12
40Kimi K2.656.6%55.2$0.9062026-10-12
41Gemini 3.6 Flashhigh56.6%61.2$0.3852026-10-12
42Claude Haiku 4.554.3%47.6$0.2252026-10-12
43Gemini 3.1 Prohigh53.8%60.1$0.3772026-10-12
44Gemini 3.5 Flash Litehigh51.8%47.9$0.0962026-10-12
45Grok 4.3high51.7%56.6$0.2642026-10-12
46Ling 3.0 Flash50.7%51.8$0.0502026-10-12
47Mercury 2.5high45.3%–$0.0682026-10-12
48Grok 4.1 Fastthinking44.8%55.6$0.0312026-10-12
49Laguna M.143.8%37.3$0.2022026-10-12
50Laguna XS.241.4%36.3$0.1132026-10-12

More knowledge benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on Public Benefits Bench v1.1

Cite as: BenchLeader, “Public Benefits Bench v1.1 leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_public_benefits.

Public Benefits Bench v1.1: questions

What does Public Benefits Bench v1.1 measure?
Questions a benefits applicant might ask about SNAP eligibility and process, graded against policy. Scores are reported in percent of tasks solved; higher is better.
What is the top score on Public Benefits Bench?
As of 6 Nov 2026, the highest score on Public Benefits Bench v1.1 is 76.9% by Claude Opus 5 (result dated 12 Oct 2026), ahead of Claude Fable 5.1 at 74.9%. Scores come from Vals AI and are re-read daily.
Which AI model leads Public Benefits Bench v1.1?
Claude Opus 5 leads Public Benefits Bench v1.1 with 76.9% as of 6 Nov 2026 (result dated 12 Oct 2026), ahead of Claude Fable 5.1 at 74.9%.
How many models have Public Benefits Bench v1.1 results?
50 model configurations have a Public Benefits Bench v1.1 result on BenchLeader, all taken from Vals AI.
Who runs Public Benefits Bench v1.1 and how often is it updated?
Public Benefits Bench v1.1 is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
Does Public Benefits Bench v1.1 count toward the BenchLeader Index?
No. Public Benefits Bench v1.1 is shown for reference but left out of the composite index.