Benchmark Almanac

On Evaluating User Simulators for Interactive Agents

Not in model cards

Models play the user in replays of 351 real sessions. The evaluation compares simulated behavior with observed behavior and measures agreement with human rankings of nine agents.

Arena and UCLA evaluated six user simulators. The headline score takes the geometric mean of ranking agreement, population fidelity, and individual session fidelity.

Made by
Arena and UCLA
Released
November 2026
Kind
Multi-turn conversation
Level
Everyday
Scored
Graded by a model
Answers
Agent task
Measures
Capability
Language
Many languages
In and out
Text → Text
Interaction
Multi-agent
Items from
Real work and use
Refresh
One fixed set
Run by
A host
6
Entries in its paper, 0 models we track

Profile-conditioned simulator + Claude Sonnet 5 tops the leaderboard with 70.0 No scores yet

Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. Where a model ran on its own with several settings, we keep its best. Agents other companies built around a model get a pale shade of that model's lab color. Gray marks blends and entries not in our model list, such as fine-tunes and older models. None of these count toward leaders. Agents built around a model, such as Profile-conditioned simulator + Claude Sonnet 5, are their own entries. We don't count them as the model's score. This view shows only models we track, each at its best.

  1. 1 Profile-conditioned simulator + Claude Sonnet 5 70.0
  2. 2 SimulatorArena + GPT-5 mini 69.0
  3. 3 Profile-conditioned simulator + GPT-5.6 Terra 64.0
  4. 4 Luna 61.0
  5. 5 τ²-bench + GPT-4.1 59.0
  6. 6 MIMESIS-9B 51.0

Source: The On Evaluating User Simulators for Interactive Agents paper

#EntryScoreSettingDateSource
1 Profile-conditioned simulator + Claude Sonnet 5 70.0 Jun 2026 Official leaderboard
2 SimulatorArena + GPT-5 mini 69.0 Aug 2025 Official leaderboard
3 Profile-conditioned simulator + GPT-5.6 Terra 64.0 Jul 2026 Official leaderboard
4 Luna 61.0 Profile-conditioned simulator Official leaderboard
5 τ²-bench + GPT-4.1 59.0 Apr 2025 Official leaderboard
6 MIMESIS-9B 51.0 Official leaderboard
#ModelScoreSettingDateSource

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 Instructions and chat

All 144 in Instructions and chat
  • IFEval In use Models answer about 500 prompts containing verifiable instructions such as word limits or required keywords. Code checks each instruction for strict and loose accuracy. Gemma 4 31B IT 98.9%
  • IFBench In use 58 new out-of-domain verifiable output constraints. Scored by programmatic constraint checks (strict or loose). Flint 89.2%
  • MultiChallenge In use Multi-turn conversations testing instruction retention, inference memory, versioned editing and self-coherence. Graded by LLM judge with rubrics. GPT-6.1 Sol 82%
  • Multi-IF In use Models follow verifiable instructions over three-turn conversations in eight languages, extending IFEval. Scored by instruction accuracy at each turn. LFM3-1.7B 89.1%
  • AdvancedIF In use Models respond to over 1,600 prompts with complex, multi-turn, and system-level instructions. A judge model checks each response against expert-written rubrics. MAI-Thinking-1 85%
  • MM-IFEval In use Multimodal models follow image-grounded instructions with exact output constraints. Checks mix rule-based verification and a judge model. GPT-4o 64.6%

Data audit

7 of 8 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 2 pass 3 notes
  • Pass Independent runs Scores from the benchmark’s paper (6 entries).
  • Pass Models scored 6 entries on its leaderboard, none a model we track.
  • Note Newest score No scores yet.
  • Note Frontier models No models scored yet.
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on arxiv.org: 4 rows for models we track and 2 other entries 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