Unspecified benchmark
LM Arena Agent Steerability is presented as a current leaderboard.
Updated Nov 13, 2026
Higher agent score ranks better on this benchmark.
Rank | Model | Agent Score | Percentile | Participants | Observations | Evidence | Evaluated |
|---|
| Rank01 | ModelGO | Agent Score19.06% (18.45% - 19.68%) | Percentile100.00% | Participants43 | Observations278.2K | EvidenceA | Evaluated |
| Rank02 | ModelOP | Agent Score14.82% (13.03% - 16.61%) | Percentile97.62% | Participants43 | Observations49.4K | EvidenceA | Evaluated |
| Rank03 | ModelAN | Agent Score14.80% (12.96% - 16.64%) | Percentile95.24% | Participants43 | Observations75.6K | EvidenceA | Evaluated |
| Rank04 | ModelGO | Agent Score13.16% (12.18% - 14.13%) | Percentile92.86% | Participants43 | Observations266K | EvidenceA | Evaluated |
| Rank05 | ModelTM | Agent Score9.28% (7.50% - 11.07%) | Percentile90.48% | Participants43 | Observations53.5K | EvidenceA | Evaluated |
| Rank06 | ModelTE | Agent Score8.78% (7.25% - 10.30%) | Percentile88.10% | Participants43 | Observations83.9K | EvidenceA | Evaluated |
| Rank07 | ModelAN | Agent Score7.88% (6.50% - 9.27%) | Percentile85.71% | Participants43 | Observations178.3K | EvidenceA | Evaluated |
| Rank08 | ModelOP | Agent Score5.54% (4.11% - 6.97%) | Percentile83.33% | Participants43 | Observations173.3K | EvidenceA | Evaluated |
| Rank09 | ModelAN | Agent Score4.38% (2.74% - 6.03%) | Percentile80.95% | Participants43 | Observations109.7K | EvidenceA | Evaluated |
| Rank10 | ModelXI | Agent Score2.85% (1.55% - 4.15%) | Percentile78.57% | Participants43 | Observations140K | EvidenceA | Evaluated |
| Rank11 | ModelAN | Agent Score2.76% (1.05% - 4.47%) | Percentile76.19% | Participants43 | Observations114.7K | EvidenceA | Evaluated |
| Rank12 | ModelME | Agent Score2.63% (1.49% - 3.77%) | Percentile73.81% | Participants43 | Observations477.4K | EvidenceA | Evaluated |
| Rank13 | ModelAN | Agent Score2.36% (1.04% - 3.69%) | Percentile71.43% | Participants43 | Observations298.1K | EvidenceA | Evaluated |
| Rank14 | ModelXI | Agent Score1.50% (0.31% - 2.69%) | Percentile69.05% | Participants43 | Observations228.7K | EvidenceA | Evaluated |
| Rank15 | ModelOP | Agent Score-0.11% (-0.75% - 0.53%) | Percentile66.67% | Participants43 | Observations620.9K | EvidenceA | Evaluated |
| Rank16 | ModelOP | Agent Score-0.52% (-1.64% - 0.60%) | Percentile64.29% | Participants43 | Observations234K | EvidenceA | Evaluated |
| Rank17 | ModelAC | Agent Score-0.78% (-1.76% - 0.21%) | Percentile61.90% | Participants43 | Observations390.1K | EvidenceA | Evaluated |
| Rank18 | ModelMA | Agent Score-0.79% (-1.65% - 0.08%) | Percentile59.52% | Participants43 | Observations562.9K | EvidenceA | Evaluated |
| Rank19 | ModelZA | Agent Score-1.03% (-2.24% - 0.18%) | Percentile57.14% | Participants43 | Observations302.5K | EvidenceA | Evaluated |
| Rank20 | ModelZA | Agent Score-1.18% (-2.32% - -0.04%) | Percentile54.76% | Participants43 | Observations501.4K | EvidenceA | Evaluated |
| Rank21 | ModelGO | Agent Score-1.26% (-2.46% - -0.06%) | Percentile52.38% | Participants43 | Observations274.9K | EvidenceA | Evaluated |
| Rank22 | ModelXA | Agent Score-1.76% (-3.61% - 0.08%) | Percentile50.00% | Participants43 | Observations123.8K | EvidenceA | Evaluated |
| Rank23 | ModelXA | Agent Score-2.27% (-4.28% - -0.25%) | Percentile47.62% | Participants43 | Observations100.5K | EvidenceA | Evaluated |
| Rank24 | ModelAC | Agent Score-2.56% (-3.66% - -1.47%) | Percentile45.24% | Participants43 | Observations384.1K | EvidenceA | Evaluated |
| Rank25 | ModelMI | Agent Score-2.57% (-4.66% - -0.47%) | Percentile42.86% | Participants43 | Observations51.1K | EvidenceA | Evaluated |
| Rank26 | ModelTE | Agent Score-3.09% (-5.07% - -1.10%) | Percentile40.48% | Participants43 | Observations53.6K | EvidenceA | Evaluated |
| Rank27 | ModelZA | Agent Score-5.01% (-6.12% - -3.89%) | Percentile38.10% | Participants43 | Observations698.6K | EvidenceA | Evaluated |
| Rank28 | ModelAC | Agent Score-6.00% (-7.42% - -4.58%) | Percentile35.71% | Participants43 | Observations266.2K | EvidenceA | Evaluated |
| Rank29 | ModelMA | Agent Score-6.39% (-9.60% - -3.18%) | Percentile33.33% | Participants43 | Observations27.1K | EvidenceA | Evaluated |
| Rank30 | ModelGO | Agent Score-7.35% (-8.80% - -5.90%) | Percentile30.95% | Participants43 | Observations303.5K | EvidenceA | Evaluated |
The leading models and scores on this benchmark.
A closer view of the leading scores on this benchmark.
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What LM Arena Agent Steerability measures and how its scores work.
LM Arena Agent Steerability is a benchmark leaderboard.
It reports Agent Score as a ratio.
Scores are shown in ratio. This benchmark is verified and has an evidence level of A.
LM Arena. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
Common questions about LM Arena Agent Steerability.
Gemini 4 Carbon is currently ranked first with 19.06%.
The current leaders are Gemini 4 Carbon (19.06%), GPT-6 Astra (14.82%), and Claude Fable 5.1 (14.80%).
Claude Haiku 5.5 has the lowest matched official input price at $0.10 input / $0.50 output per 1M tokens.
The fastest matched records are MiniMax M3 (156.78 tok/s via MiniMax), Mistral Large 4 (107.24 tok/s via Mistral AI), and Gemini 3.1 Pro (90.00 tok/s via Google).
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.
It reports Agent Score as a ratio.
Yes. Higher values rank better for this benchmark.
43 model results are currently shown.
No. This benchmark is shown for reference but does not contribute to the overall score.
Ranking basisThis lm arena agent steerability AI model leaderboard uses descending agent score in the benchmark original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
Selection summary
Gemini 4 Carbon currently leads LM Arena Agent Steerability with 19.06%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price, and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.