All General Coding Reasoning Multimodal Safety Specialized Agentic Language Long Context Speed Text-to-Image Image Editing Text-to-Video Image-to-Video
BFCL83 BrowseComp OSWorld-Verified TAU-bench18 TAU-bench Airline22 TAU-bench Banking Knowledge27 TAU-bench Retail23 TAU-bench Telecom18 Tau Banking (Artificial Analysis)209 Tau2 Telecom (Artificial Analysis)439 Terminal-Bench 2.1241 Terminal-Bench Hard431 ToolBench WebArena

TAU-bench Banking Knowledge

Agentic

Source ↗ Higher is better

Rank Model Provider Score
1 Qwen 3.8 Max Alibaba
55.2%
2 Anthropic: Claude Opus 5 Anthropic
48.7%
3 SpaceXAI: Grok 4.5 SpaceXAI
47.9%
4 OpenAI: GPT-5.6 Sol OpenAI
46.9%
5 OpenAI: GPT-5.5 OpenAI
44.6%
6 Meta: Muse Spark 1.1 Meta
40.5%
7 Anthropic: Claude Opus 4.7 Anthropic
40.2%
8 Anthropic: Claude Opus 4.8 Anthropic
39.7%
9 Anthropic: Claude Fable 5 Anthropic
39.7%
10 OpenAI: GPT-5.4 OpenAI
39.4%
11 Z.ai: GLM 5.2 Z.ai
37.1%
12 MoonshotAI: Kimi K3 Moonshot AI
37.1%
13 OpenAI: GPT-5.2 OpenAI
32.2%
14 Distyl ButtonAgent Distyl AI
31.2%
15 Gemini 3 Flash Google
27.3%
16 Anthropic: Claude Opus 4.6 Anthropic
27.3%
17 Google: Gemini 3.1 Pro Preview Google
26.0%
18 Anthropic: Claude Sonnet 4.5 Anthropic
25.3%
19 Thinking Machines: Inkling Thinking Machines
25.0%
20 Anthropic: Claude Opus 4.5 Anthropic
24.7%
21 gemini-3-pro Google
18.0%
22 Grok 4.2 SpaceXAI
18.0%
23 Grok 4 fast SpaceXAI
15.7%
24 Google: Gemini 2.5 ProRetiring Google
13.7%
25 Grok 4.1 fast SpaceXAI
13.1%
26 Qwen: Qwen3.5 397B A17B Alibaba
9.8%
27 Z.ai: GLM 5 Z.ai
9.8%

Access this data

JSON GET /api/v1 
[
  {
    "rank": 1,
    "model": "Qwen 3.8 Max",
    "provider": "Alibaba",
    "score": "55.15463917525773"
  },
  {
    "rank": 2,
    "model": "Anthropic: Claude Opus 5",
    "provider": "Anthropic",
    "score": "48.71134020618557"
  },
  {
    "rank": 3,
    "model": "SpaceXAI: Grok 4.5",
    "provider": "SpaceXAI",
    "score": "47.93814432989691"
  }
]
Curl 
$ curl -s https://rightrank.ai/api/v1/rankings/tau-bench-banking-knowledge.json | jq '.[:3]'
CLI brew install rightrank 
$ rightrank top tau-bench-banking-knowledge --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category tau-bench-banking-knowledge
$ rightrank value tau-bench-banking-knowledge  # best performance per dollar
Claude Code Skill Skill 
# In Claude Code, use the RightRank skill:
/rightrank best model for tau-bench-banking-knowledge
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on tau-bench-banking-knowledge