Benchmark Almanac

CounselBench-Adv

GPT-4o Aug 2024 0.37 leads since May 2025
Not in model cards

Models answer 120 questions written by mental health professionals. Each question targets one of six problems identified in earlier counseling responses.

Slingshot AI tested ten models and reports the fraction of answers that exhibit the targeted problem.

Released
May 2025
Kind
Wellbeing and sycophancy
Level
Expert
Scored
Graded by a model
Answers
Open-ended
Tasks
All public
Measures
Propensity
Language
English
In and out
Text → Text
Interaction
Single turn
Items from
Written by experts
Refresh
One fixed set
Run by
Anyone
Occupations
Healthcare practitioners and technical
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
September 2026
Scores refreshed
November 20th, 2026
10
Models scored on its own leaderboard

GPT-4o leads with 0.37 among models released by September 2026

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.

The newest model with a score here came out in September 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

  1. 1 GPT-4o Aug 2024 0.37
  2. 2 GPT-6 Luna 0.41
  3. 3 GPT-6 Astra 0.43
  4. 4 Claude Opus 5.5 0.49
  5. 5 Qwen3.5-122B-A10B 0.49
  6. 6 Qwen3.5-27B 0.49
  7. 7 GPT-5 0.49
  8. 8 Gemini 2.5 Pro Jun 2025 0.54
  9. 9 Claude Sonnet 4.5 0.57
  10. 10 Gemini 4 Argon 0.57

Source: Official CounselBench-Adv leaderboard

#ModelScoreSettingDateLinkSource
1 GPT-4o Aug 2024 0.37 Aug 2024 github.com Official leaderboard
2 GPT-6 Luna 0.41 Sep 2026 github.com Official leaderboard
3 GPT-6 Astra 0.43 Sep 2026 github.com Official leaderboard
4 Claude Opus 5.5 0.49 Sep 2026 github.com Official leaderboard
5 Qwen3.5-122B-A10B 0.49 Feb 2026 github.com Official leaderboard
6 Qwen3.5-27B 0.49 Feb 2026 github.com Official leaderboard
7 GPT-5 0.49 Aug 2025 github.com Official leaderboard
8 Gemini 2.5 Pro Jun 2025 0.54 Jun 2025 github.com Official leaderboard
9 Claude Sonnet 4.5 0.57 Sep 2025 github.com Official leaderboard
10 Gemini 4 Argon 0.57 Sep 2026 github.com Official leaderboard

Rankings by part

We read 6 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Apathetic, Assumptions and Symptoms. GPT-4o leads 3 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Apathetic

10 models we track have a score here.

  1. 1 Gemini 2.5 Pro Jun 2025 ↑7 0.00
  2. 2 Qwen3.5-122B-A10B ↑3 0.00
  3. 3 Qwen3.5-27B ↑3 0.00
  4. 4 GPT-6 Astra ↓1 0.00
  5. 5 Claude Opus 5.5 ↓1 0.00
  6. 6 Gemini 4 Argon ↑4 0.00
  7. 7 GPT-5 0.02
  8. 8 GPT-6 Luna ↓6 0.03
  9. 9 GPT-4o Aug 2024 ↓8 0.07
  10. 10 Claude Sonnet 4.5 ↓1 0.07

Number one over time

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

OpenAI
Own leaderboard Not tested since Sep 2026
GPT-4o 0.37

Source: Official CounselBench-Adv 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
0.40 0.50 0.60 2025 2026 0.40 0.50 0.60 2025 2026

Source: Official CounselBench-Adv 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 Safety and alignment

All 636 in Safety and alignment
  • FORTRESS In use Paired adversarial and benign national-security misuse prompts. Rubric-graded risk score on adversarial and over-refusal score on benign prompts. Claude Opus 5.5 5.78%
  • StrongREJECT In use Forbidden-prompt jailbreak benchmark with a rubric grader that scores how useful a non-refusal actually is. Inkling 98.6%
  • Gray Swan IPI Arena benchmark In use Indirect prompt-injection attacks from public red-teaming competitions replayed on tool, coding and browser agents. Attack success rate at k attempts. Gemini 4 Carbon 0%
  • Covert misbehavior (Thinking Machines) In use Agentic tasks that check whether a model misbehaves and conceals it. Scored by the rate of covert misbehavior. Inkling 2 Mini 8.30%
  • Destructive actions (Thinking Machines) In use Agentic tasks that check whether a model takes destructive actions it was not authorized to take. Scored by the rate of destructive actions. Inkling 2 Small 4.70%
  • Misbehavior (Thinking Machines) In use Agentic tasks that check whether a model fabricates results, ignores instructions or hides what it did when a task is hard. Scored by the rate of misbehavior. Inkling 2 Mini 9.80%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (10 entries).
  • Pass Models scored 10 models we track.
  • Pass Newest score The newest is from September 2026, for Gemini 4 Argon.
  • Pass Frontier models 3 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on mheval.slingshotai.com: 10 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