All General Coding Reasoning Multimodal Safety Specialized Agentic Language Long Context Speed Text-to-Image Image Editing Text-to-Video Image-to-Video
AGIEval ARC Artificial Analysis Intelligence Index689 Chatbot Arena ELO282 GDPval-AA HellaSwag Humanity's Last Exam661 IFBench449 LiveBench MMLU MMLU-Pro349 MMLU-Pro (Open LLM Leaderboard)2839 WinoGrande

General

Composite performance across 6 benchmarks · Normalized 0–100 · At least 3 scores required · Ranked by average performance · Top 50 of 518 eligible models

# Model Provider AA Index Arena ELO Humanity's ... IFBench MMLU-Pro MMLU-Pro Open Avg Coverage Price Value
1 Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) Anthropic 49.6 — 55.5% 63.5% — —
82.5
3/6 $60.00/M 1.4
2 GPT-5.6 Sol (max) OpenAI 47.0 — 49.5% 72.7% — —
82.0
3/6 $24.00/M 3.4
3 GPT-5.6 Sol (xhigh) OpenAI 44.0 — 47.3% 71.0% — —
78.1
3/6 $24.00/M 3.3
4 GPT-5.5 (xhigh) OpenAI 38.4 — 45.8% 75.9% — —
76.1
3/6 $35.00/M 2.2
5 GPT-5.6 Sol (high) OpenAI 42.3 — 46.0% 69.2% — —
75.5
3/6 $24.00/M 3.1
6 GPT-5.6 Terra (max) OpenAI 42.1 — 42.9% 71.2% — —
74.6
3/6 $14.00/M 5.3
7 GPT-5.4 (xhigh) OpenAI 39.0 — 43.7% 73.9% — —
74.4
3/6 $17.50/M 4.3
8 GPT-5.2 (xhigh) OpenAI 30.4 — 37.7% 75.4% 87.4% —
74.3
4/6 $15.75/M 4.7
9 Claude Opus 4.8 (Adaptive Reasoning, Max Effort) Anthropic 41.8 — 48.7% 62.2% — —
73.5
3/6 $30.00/M 2.4
10 Gemini 3 Flash Preview (Reasoning) Google 26.3 — 36.6% 78.0% 89.0% —
73.3
4/6 $3.50/M 20.9
11 Gemini 3 Pro Preview (high) Google 28.0 — 39.7% 70.4% 89.8% —
72.9
4/6 $14.00/M 5.2
12 GPT-5.5 (high) OpenAI 37.0 — 45.0% 71.6% — —
72.9
3/6 $35.00/M 2.1
13 GPT-5.1 (high) OpenAI 24.7 1,446 28.5% 72.9% 87.0% —
72.2
5/6 $11.25/M 6.4
14 Google: Gemini 3.1 Pro Preview Google 29.7 — 47.0% 77.1% — —
72.1
3/6 $14.00/M 5.1
15 GPT-5.6 Sol (medium) OpenAI 39.2 — 42.2% 69.6% — —
71.7
3/6 $24.00/M 3.0
16 Gemini 3.5 Flash (high) Google 32.6 — 42.7% 76.3% — —
71.1
3/6 $10.50/M 6.8
17 GPT-5.3 Codex (xhigh) OpenAI 32.5 — 42.5% 75.4% — —
70.5
3/6 $15.75/M 4.5
18 GPT-5 (high) OpenAI 23.0 1,407 28.5% 73.1% 87.1% —
70.4
5/6 $11.25/M 6.3
19 Gemini 3.5 Flash (medium) Google 33.6 — 41.3% 74.6% — —
70.1
3/6 $10.50/M 6.7
20 MiniMax: MiniMax M3 MiniMax 29.2 — 39.0% 82.9% — —
70.1
3/6 $1.50/M 46.7
21 Qwen: Qwen3.7 Max Alibaba 29.5 — 40.5% 80.5% — —
70.0
3/6 $5.90/M 11.9
22 GLM-5.2 (max) Z.ai 33.7 — 41.1% 73.3% — —
69.5
3/6 $7.04/M 9.9
23 GPT-5.6 Terra (xhigh) OpenAI 38.0 — 41.9% 66.3% — —
69.3
3/6 $14.00/M 4.9
24 GPT-5.5 (medium) OpenAI 33.8 — 42.4% 71.0% — —
69.2
3/6 $35.00/M 2.0
25 Muse Spark Meta 31.3 — 40.7% 75.9% — —
69.0
3/6 — —
26 Claude Opus 4.7 (Adaptive Reasoning, Max Effort) Anthropic 40.7 — 42.3% 58.6% — —
67.5
3/6 $30.00/M 2.3
27 DeepSeek V4 Pro 0424 (Reasoning, Max Effort) DeepSeek 30.4 — 37.5% 76.5% — —
67.0
3/6 $1.30/M 51.3
28 GPT-5 Codex (high) OpenAI 24.9 — 27.8% 74.1% 86.5% —
66.9
4/6 $11.25/M 5.9
29 Grok 4.3 (high) SpaceXAI 24.9 — 37.2% 81.3% — —
65.7
3/6 $3.75/M 17.5
30 GPT-5.2 Codex (xhigh) OpenAI 28.5 — 35.7% 77.6% — —
65.4
3/6 $15.75/M 4.2
31 GPT-5.6 Sol (low) OpenAI 33.5 — 39.4% 66.5% — —
65.3
3/6 $24.00/M 2.7
32 Claude Opus 4.5 (Reasoning) Anthropic 29.1 — 30.1% 58.0% 89.5% —
65.0
4/6 $30.00/M 2.2
33 MiMo-V2.5-Pro (Reasoning) Xiaomi 26.0 — 35.7% 79.9% — —
64.9
3/6 $1.30/M 49.7
34 GPT-5 mini (high) OpenAI 16.8 1,375 21.5% 75.4% 83.7% —
64.7
5/6 $2.25/M 28.8
35 Kimi K2.6 (Reasoning) Moonshot AI 27.0 — 37.5% 76.0% — —
64.7
3/6 $4.06/M 15.9
36 Grok 4.20 0309 v2 (Reasoning) SpaceXAI 25.7 — 34.5% 81.2% — —
64.7
3/6 $3.75/M 17.2
37 Grok 4 SpaceXAI 22.5 1,409 26.7% 53.7% 86.6% —
64.2
5/6 $18.00/M 3.6
38 GPT-5.6 Terra (high) OpenAI 34.2 — 38.5% 64.4% — —
64.2
3/6 $14.00/M 4.6
39 Grok 4.20 0309 (Reasoning) SpaceXAI 25.2 — 32.4% 82.9% — —
64.0
3/6 $8.00/M 8.0
40 GPT-5.2 (medium) OpenAI 26.5 — 26.7% 65.2% 85.9% —
63.9
4/6 $15.75/M 4.1
41 GPT-5.1 Codex (high) OpenAI 23.7 — 25.7% 70.0% 86.0% —
63.9
4/6 $11.25/M 5.7
42 GPT-5 (medium) OpenAI 22.9 — 25.4% 70.6% 86.7% —
63.8
4/6 $11.25/M 5.7
43 Qwen: Qwen3.7 Plus Alibaba 25.2 — 35.6% 78.0% — —
63.5
3/6 $1.60/M 39.7
44 DeepSeek V4 Pro 0424 (Reasoning, High Effort) DeepSeek 30.1 — 35.2% 71.3% — —
63.1
3/6 $1.30/M 48.4
45 GLM-4.7 (Reasoning) Z.ai 22.2 — 27.4% 67.9% 85.6% —
63.0
4/6 $2.80/M 22.5
46 o1-preview OpenAI 11.4 1,353 — — 84.8% —
63.0
3/6 $82.50/M 0.8
47 DeepSeek V4 Flash 0420 (Reasoning, Max Effort) DeepSeek 24.2 — 34.8% 79.2% — —
63.0
3/6 $0.35/M 182.0
48 KAT-Coder-Pro V1 Kwaipilot 18.6 — 33.6% 68.4% 81.3% —
62.9
4/6 — —
49 Grok 4.3 (medium) SpaceXAI 24.8 — 30.0% 83.3% — —
62.6
3/6 $3.75/M 16.7
50 Qwen: Qwen3.6 Max PreviewHistorical Alibaba 28.4 — 30.8% 76.6% — —
62.1
3/6 $9.10/M 6.8

Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback) leads on raw performance, but DeepSeek V4 Flash 0420 (Reasoning, Max Effort) delivers 12900% more performance per dollar. That extra 19.5 points costs $3.06/M per point.

Access this data

JSON GET /api/v1 
[
  {
    "rank": 1,
    "model": "Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)",
    "provider": "Anthropic",
    "avg_score": 82.5,
    "cost": 60.0,
    "cost_unit": "million_tokens",
    "value": 1.4
  },
  {
    "rank": 2,
    "model": "GPT-5.6 Sol (max)",
    "provider": "OpenAI",
    "avg_score": 82.0,
    "cost": 24.0,
    "cost_unit": "million_tokens",
    "value": 3.4
  },
  {
    "rank": 3,
    "model": "GPT-5.6 Sol (xhigh)",
    "provider": "OpenAI",
    "avg_score": 78.1,
    "cost": 24.0,
    "cost_unit": "million_tokens",
    "value": 3.3
  }
]
Curl 
$ curl -s https://rightrank.ai/api/v1/rankings/general.json | jq '.[:3]'
CLI brew install rightrank 
$ rightrank top general --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category general
$ rightrank value general  # best performance per dollar
Claude Code Skill Skill 
# In Claude Code, use the RightRank skill:
/rightrank best model for general
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on general