Tau2 Telecom (Artificial Analysis)

Agentic

Source ↗ Higher is better

Rank	Model	Provider	Score
1	GLM-5.2 (max)	Z.ai	99.1%
2	JT-35B-Flash	China Mobile	99.1%
3	GLM-4.7-Flash (Reasoning)	Z.ai	98.8%
4	Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)	Anthropic	98.5%
5	StepFun: Step 3.7 Flash	Stepfun	98.5%
6	GLM 5V Turbo (Reasoning)	Z.ai	98.5%
7	Z.ai: GLM 5 TurboRetiring	Z.ai	98.5%
8	GLM-5 (Reasoning)	Z.ai	98.2%
9	Qwen: Qwen3.6 Plus	Alibaba	97.7%
10	GLM-5.1 (Reasoning)	Z.ai	97.7%
11	Grok 4.3 (high)	SpaceXAI	97.7%
12	GLM-5 (Non-reasoning)	Z.ai	97.4%
13	GLM-5.1 (Non-reasoning)	Z.ai	97.1%
14	Grok 4.20 0309 (Reasoning)	SpaceXAI	96.5%
15	DeepSeek V4 Pro 0424 (Reasoning, Max Effort)	DeepSeek	96.2%
16	Kimi K2.6 (Reasoning)	Moonshot AI	95.9%
17	Qwen: Qwen3.6 Max PreviewHistorical	Alibaba	95.9%
18	Kimi K2.5 (Reasoning)	Moonshot AI	95.9%
19	GLM-4.7 (Reasoning)	Z.ai	95.9%
20	Google: Gemini 3.1 Pro Preview	Google	95.6%
21	Gemini 3.5 Flash (medium)	Google	95.6%
22	Qwen3.5 397B A17B (Reasoning)	Alibaba	95.6%
23	DeepSeek V4 Flash 0420 (Reasoning, High Effort)	DeepSeek	95.6%
24	MiniMax: MiniMax M2.5	MiniMax	95.3%
25	Gemini 3.5 Flash (high)	Google	95.3%
26	Qwen3.6 35B A3B (Reasoning)	Alibaba	95.3%
27	DeepSeek V4 Flash 0420 (Reasoning, Max Effort)	DeepSeek	95.0%
28	MiMo-V2-Pro	Xiaomi	95.0%
29	MiMo-V2-Flash (Reasoning)	Xiaomi	95.0%
30	Qwen: Qwen3.7 Max	Alibaba	94.7%
31	Claude Opus 4.8 (Adaptive Reasoning, Max Effort)	Anthropic	94.4%
32	DeepSeek V4 Flash 0420 (Non-reasoning)	DeepSeek	94.4%
33	StepFun: Step 3.5 Flash	Stepfun	94.4%
34	DeepSeek V4 Pro 0424 (Reasoning, High Effort)	DeepSeek	94.2%
35	Mistral: Mistral Medium 3.5	Mistral AI	94.2%
36	Qwen3.6 27B (Reasoning)	Alibaba	94.2%
37	MiMo-V2.5-Pro (Reasoning)	Xiaomi	94.2%
38	GLM-4.7 (Non-reasoning)	Z.ai	94.2%
39	Kimi K2.6 (Non-reasoning)	Moonshot AI	93.9%
40	Qwen3.5 27B (Reasoning)	Alibaba	93.9%
41	GPT-5.5 (xhigh)	OpenAI	93.9%
42	Qwen3.5 122B A10B (Reasoning)	Alibaba	93.6%
43	Qwen3.6 27B (Non-reasoning)	Alibaba	93.6%
44	Tri-21B-think Preview	Trillion Labs	93.3%
45	MiMo-V2-Flash (Feb 2026)	Xiaomi	93.3%
46	grok-4-1-fast-reasoning	SpaceXAI	93.3%
47	JT-MINI	China Mobile	93.0%
48	Grok 4.20 0309 v2 (Reasoning)	SpaceXAI	93.0%
49	MoonshotAI: Kimi K2 Thinking	Moonshot AI	93.0%
50	GPT-5.5 (high)	OpenAI	93.0%

Access this data

JSON GET /api/v1

[
  {
    "rank": 1,
    "model": "GLM-5.2 (max)",
    "provider": "Z.ai",
    "score": "99.1228070175439"
  },
  {
    "rank": 2,
    "model": "JT-35B-Flash",
    "provider": "China Mobile",
    "score": "99.1228070175439"
  },
  {
    "rank": 3,
    "model": "GLM-4.7-Flash (Reasoning)",
    "provider": "Z.ai",
    "score": "98.8304093567251"
  }
]

Curl

$ curl -s https://rightrank.ai/api/v1/rankings/tau2-telecom-artificial-analysis.json | jq '.[:3]'

CLI brew install rightrank

$ rightrank top tau2-telecom-artificial-analysis --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category tau2-telecom-artificial-analysis
$ rightrank value tau2-telecom-artificial-analysis  # best performance per dollar

Claude Code Skill Skill

# In Claude Code, use the RightRank skill:
/rightrank best model for tau2-telecom-artificial-analysis
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on tau2-telecom-artificial-analysis