BasedAGIBasedAGI
Menu
Rankings live

customer_experience

Ticket thread summary

Summarize long ticket threads into status, attempts, and next steps.

#1 Recommendation

gemini-2.5-pro

Strong on Galileo Agent Leaderboard v2 Avg AC (59%) and FACTS Benchmark Suite facts_grounding_score_pct (100%)

external/google/gemini-2-5-pro

21.6%

Score

33.6%

Confidence

Limited benchmark evidence for this use case.

49 ranked models with average evidence of 14.0 points. Rankings may shift as more benchmark data is ingested.

Ranked Models

30

Evidence Quality

80%

Scoring

Benchmark-backed

Top Signal

Galileo Agent Leaderboard v2: Avg AC

All Ranked Models

Max params:
Min confidence:
30 of 30
RankModelScore
#1gemini-2.5-pro

Strong on Galileo Agent Leaderboard v2 Avg AC (59%) and FACTS Benchmark Suite facts_grounding_score_pct (100%)

21.6%
#2gemini-3-pro-preview

Strong on FACTS Benchmark Suite facts_grounding_score_pct (88%) and Vals Finance Agent overall_accuracy_pct (87%)

20.9%
#3gpt-4.1-20250414

Strong on Galileo Agent Leaderboard v2 Avg AC (100%) and Vectara HHEM Leaderboard overall_hallucination_error_pct (82%)

20.5%
#4Grok-4-0709
17.1%
#5claude-sonnet-4-20250514
16.1%
#6anthropic/claude-sonnet-4.6
16.0%
#7gpt-5-mini-2025-08-07
15.7%
#8gemini-2.5-flash
15.2%
#9gpt-5-2025-08-07
15.0%
#10google/gemini-3.1-pro-preview
15.0%
#11openai/gpt-5.4-2026-03-05
14.3%
#12gpt-5.1-2025-11-13
13.0%
#13qwen-2.5-72b-instruct
12.7%
#14claude-opus-4-5-20251101
12.4%
#15gpt-4o
12.2%
#16gemini-3-flash-preview
11.6%
#17google/gemini-3.1-flash-lite-preview
11.4%
#18xai-org/grok-4-fast-reasoning
11.2%
#20gpt-5.2-2025-12-11
11.0%
#25anthropic/claude-opus-4-6-thinking
10.6%
#26xai-org/grok-4-1-fast-reasoning
10.4%
#34anthropic/claude-opus-4-5-20251101-thinking
9.9%
#40Qwen3-Embedding-4B
9.8%
#47gpt-4.1-mini-20250414
9.5%
#50kimi/kimi-k2.5-thinking
9.5%
#56anthropic/claude-sonnet-4-5-20250929-thinking
9.4%
#75gpt-4o-2024-08-06
8.7%
#83anthropic/claude-opus-4-1-20250805
8.6%
#91deepseek-v3
8.5%
#119mistralai/mistral-large-2512
8.1%

Compare Models

Model A leads by +0.7%

Shareable Link →

Model A

gemini-2.5-pro

external/google/gemini-2-5-pro

21.6%

Rank #1

Confidence 33.6%23 evidence pts

Galileo Agent Leaderboard v2: Avg AC

Value 58.7% · Conf 100.0% · Weight 2.2%

galileo_agent_v2.avg_ac (Mar 12, 2026)

FACTS Benchmark Suite: facts_grounding_score_pct

Value 100.0% · Conf 100.0% · Weight 1.8%

facts_benchmark_suite.facts_grounding_score_pct (Mar 12, 2026)

Vectara HHEM Leaderboard: overall_hallucination_error_pct

Value 76.0% · Conf 100.0% · Weight 1.7%

vectara_hhem_leaderboard.overall_hallucination_error_pct (Mar 12, 2026)

SciArena Leaderboard: rating_elo

Value 70.7% · Conf 100.0% · Weight 1.4%

sciarena_leaderboard.rating_elo (Mar 12, 2026)

Model B

gemini-3-pro-preview

external/google/gemini-3-pro-preview

20.9%

Rank #2

Confidence 28.9%22 evidence pts

FACTS Benchmark Suite: facts_grounding_score_pct

Value 88.3% · Conf 100.0% · Weight 1.6%

facts_benchmark_suite.facts_grounding_score_pct (Mar 12, 2026)

Vals Finance Agent: overall_accuracy_pct

Value 87.0% · Conf 100.0% · Weight 1.6%

vals_finance_agent.overall_accuracy_pct (Mar 12, 2026)

SciArena Leaderboard: rating_elo

Value 78.8% · Conf 100.0% · Weight 1.6%

sciarena_leaderboard.rating_elo (Mar 12, 2026)

FACTS Benchmark Suite: average_score_pct

Value 100.0% · Conf 100.0% · Weight 1.6%

facts_benchmark_suite.average_score_pct (Mar 12, 2026)

Ranking Diagnostics & Missing Models

Source Lift

Ranked

49

Sources

8

Quality

Insufficient

Vals Legal Bench

vals_legal_bench

36 rows

0.4% avg lift

Vals MedQA

vals_medqa

36 rows

0.4% avg lift

Vals Tax Eval v2

vals_tax_eval_v2

36 rows

0.4% avg lift

Vals LiveCodeBench

vals_lcb

34 rows

0.3% avg lift

Missing Strong Models

google/gemini-2.0-flash-001

external/google/gemini-2-0-flash-001

Rank #56

10.3%

Thin evidence after weighting

Llama-4-Scout-17B-16E-Instruct

external/meta/llama-4-scout-17b-16e-instruct

Rank #73

8.4%

Thin evidence after weighting
Taxonomy Details

Core Tasks

task.summarize_ticket_threadtask.timeline_extraction

Required Modes

mode.long_context

Domains

domain.customer_support

Related Use Cases