LLMBenchmark.ai

Enter opens the highlighted result. Esc closes.

Refreshed daily from 18 public sources · data from October 20, 2026

LLM leaderboard and AI benchmarks, October 2026

As of October 20, 2026, Anthropic Claude Opus 5.5 leads the LLMBenchmark.ai leaderboard of 213 ranked models with 57.6 on the Artificial Analysis Intelligence Index, and Anthropic Claude Opus 5.5 leads coding with 92.2 on the Vibe Coding Index from 2 benchmarks. Every number links to the board it was read from and the day it was read.

  • 356models tracked
  • 213ranked overall
  • 190public benchmarks
  • 9180head-to-head pages

The whole market: 356 models from 57 makersPoint at a dot

Claude Opus 5.51 Anthropic · #1 of 213 overall · coding 92.2 · $20.00 / M out 57.6
OpenAI Qwen Google Mistral AI Z.ai Anthropic DeepSeek Meta MiniMax NVIDIA xAI Amazon Perplexity Xiaomi Inception inclusionAI StepFun Upstage Thinking Machines
  • Dot size: overall index
  • Hollow ring: not ranked yet
  • Colour and bubble: the maker

Overall index: Artificial Analysis. Catalog and prices: OpenRouter. Coding: our Vibe Coding Index. Fetched October 20, 2026.

Leaders on October 20

Current leaders

The top three overall today, each with its place on every category board, among every model measured there. Full leaderboard

  1. 1

    Claude Opus 5.5

    Anthropic

    57.6Overall index, leads by 1.6

    • 1 Coding 92.2 #1/136
    • 1 Math 100.0% #1/116
    • 4 Web design 1336 #4/131
    • 35 Reasoning 90.6% #35/203
    • Agentic Not measured yet
    • Knowledge Not measured yet

    $20.00 per M output, 1M context

  2. 2

    Claude Sonnet 5.5

    Anthropic

    56.0Overall index, 1.6 behind the leader

    • 1 Math 100.0% #1/116
    • 3 Coding 88.9 #3/136
    • 3 Reasoning 95.6% #3/203
    • 6 Web design 1326 #6/131
    • Agentic Not measured yet
    • Knowledge Not measured yet

    $10.00 per M output, 1M context

  3. 3

    Claude Fable 5.1

    Anthropic

    53.4Overall index, 4.2 behind the leader

    • 1 Agentic 57.9 #1/87
    • 1 Math 100.0% #1/116
    • 1 Knowledge 92.4% #1/147
    • 2 Coding 89.1 #2/136
    • 10 Web design 1313 #10/131
    • 15 Reasoning 93.4% #15/203

    $50.00 per M output, 1M context

Overall leaderboard

Ordered by the Artificial Analysis Intelligence Index. The tabs re-rank by one category and keep the models measured on it. Press + to compare; every pairing opens from the bar at the bottom. Open as its own page · Compare models

20 of 213 shown

JSON
LLMs ranked overall, one column per category
Rank
1
Claude Opus 5.5 Anthropic
57.6
92.2 No data 90.6%* 100.0% No data 1336 $20.00
2
Claude Sonnet 5.5 Anthropic
56.0
88.9 No data 95.6%* 100.0% No data 1326 $10.00
3
Claude Fable 5.1 Anthropic
53.4
89.1 57.9 93.4% 100.0% 92.4% 1313 $50.00
4
GPT-6 Astra OpenAI
52.7
88.7 51.0 95.8%* 100.0% No data 1338 $50.00
5
Gemini 4 Argon Google
52.6
80.5 No data 96.4%* 100.0% No data 1315 $10.00
6
GPT-6.1 Sol OpenAI
51.8
88.7 No data 95.4%* 100.0% No data 1320 $10.00
7
Claude Opus 5 Anthropic
50.8
81.8 56.5 93.4% 98.9% 91.6% 1310 $25.00
8
Claude Fable 5 Anthropic
49.6
82.1 50.7 93.2% 100.0% 91.5% 1296 $50.00
9
Muse Spark 1.3 Meta
48.1
79.7 55.5 94.1%* 99.2% No data 1356 $4.25
10
GPT-6 Sol OpenAI
47.6
82.8 No data 94.3%* 100.0% No data 1292 $10.00
11
GPT-5.6 Sol OpenAI
47.0
76.7 50.2 95.2% 100.0% 89.1% 1321 $20.00
12
Grok 4.7 xAI
46.4
75.0 No data 92.7%* 98.1% No data 1253 $6.00
13
MiMo-V2.6-Pro Xiaomi
46.3
72.6 No data No data No data No data 1308 $0.87
14
Qwen3.8 Max (0902) Qwen
45.4
63.5 56.0 93.7% 100.0% 88.6% 1276 $6.00
15
GLM 5.3 Z.ai
44.8
73.1 53.1 88.1% 91.1% 86.8% 1295 $1.40
16
Grok 4.6 xAI
44.3
76.4 53.0 94.7% 99.2% 89.4% 1289 $6.00
17
Step 5 Preview StepFun
43.7
71.4 No data No data No data No data No data $2.70
18
Kimi K3 Moonshot AI
43.6
61.8 50.0 92.9% 97.2% 88.0% 1335 $11.936
19
Claude Haiku 5.5 Anthropic
43.4
75.9 No data 89.6%* 98.9% No data 1263 $0.50
20
GPT-5.6 Terra OpenAI
42.1
69.0 43.2 90.9% 99.7% 86.7% 1269 $12.00

First on that board Top 5% Top 15% Top 35% of the models it measured

Find your model

Three questions, five answers. Every pick is ranked by a public board, and a model with no number for what you chose is left out and counted, not guessed.

AA Intelligence Index

  1. 1 Claude Opus 5.5Anthropic 57.6 $20.00 / M out · 1M context · 96 tok/s
  2. 2 Claude Sonnet 5.5Anthropic 56.0 $10.00 / M out · 1M context · 137 tok/s
  3. 3 Claude Fable 5.1Anthropic 53.4 $50.00 / M out · 1M context · 73 tok/s
  4. 4 GPT-6 AstraOpenAI 52.7 $50.00 / M out · 1.05M context · 69 tok/s
  5. 5 Gemini 4 ArgonGoogle 52.6 $10.00 / M out · 1.05M context · 81 tok/s

224 models fit; 11 of them have no number for this and are left out.

Quality against price

Every paid model on the overall index. Up is better, left is cheaper; the dashed line is the best you can get at each price.

0102030405060 $0.1$0.3$1$3$10$30$100 median $1.95 median 18.5 Better and cheaper than the median: 35Better, dearer: 70Weaker, cheaper: 71Weaker and dearer than the median: 35 Claude Opus 5.5 (Anthropic): overall 57.6, $20.00 per million output tokens Claude Sonnet 5.5 (Anthropic): overall 56.0, $10.00 per million output tokens Claude Fable 5.1 (Anthropic): overall 53.4, $50.00 per million output tokens GPT-6 Astra (OpenAI): overall 52.7, $50.00 per million output tokens Gemini 4 Argon (Google): overall 52.6, $10.00 per million output tokens GPT-6.1 Sol (OpenAI): overall 51.8, $10.00 per million output tokens Claude Opus 5 (Anthropic): overall 50.8, $25.00 per million output tokens Claude Fable 5 (Anthropic): overall 49.6, $50.00 per million output tokens Muse Spark 1.3 (Meta): overall 48.1, $4.25 per million output tokens GPT-6 Sol (OpenAI): overall 47.6, $10.00 per million output tokens GPT-5.6 Sol (OpenAI): overall 47.0, $20.00 per million output tokens Grok 4.7 (xAI): overall 46.4, $6.00 per million output tokens MiMo-V2.6-Pro (Xiaomi): overall 46.3, $0.87 per million output tokens Qwen3.8 Max (0902) (Qwen): overall 45.4, $6.00 per million output tokens GLM 5.3 (Z.ai): overall 44.8, $1.40 per million output tokens Grok 4.6 (xAI): overall 44.3, $6.00 per million output tokens Step 5 Preview (StepFun): overall 43.7, $2.70 per million output tokens Kimi K3 (Moonshot AI): overall 43.6, $11.936 per million output tokens Claude Haiku 5.5 (Anthropic): overall 43.4, $0.50 per million output tokens GPT-5.6 Terra (OpenAI): overall 42.1, $12.00 per million output tokens GLM 5.3 Flash (Z.ai): overall 41.8, $0.16 per million output tokens Claude Opus 4.8 (Anthropic): overall 41.8, $25.00 per million output tokens Ling 3.1 Flash (inclusionAI): overall 41.1, $0.225 per million output tokens Gemini 3.8 Flash (Google): overall 40.9, $4.125 per million output tokens Claude Opus 4.7 (Anthropic): overall 40.7, $25.00 per million output tokens Qwen3.8 2.4T A95B (Qwen): overall 39.9, $6.00 per million output tokens Gemini 3.7 Flash (Google): overall 39.6, $3.75 per million output tokens Muse Spark 1.2 (Meta): overall 39.6, $4.25 per million output tokens DeepSeek V4.1 Flash 1017 (DeepSeek): overall 39.5, $0.60 per million output tokens GPT-5.4 (OpenAI): overall 39.0, $15.00 per million output tokens Grok 4.5 (xAI): overall 38.8, $6.00 per million output tokens GPT-5.5 (OpenAI): overall 38.4, $30.00 per million output tokens Mistral Large 4 (Mistral AI): overall 38.4, $2.09 per million output tokens Claude Sonnet 5 (Anthropic): overall 38.2, $10.00 per million output tokens GPT-6 Luna (OpenAI): overall 38.1, $0.50 per million output tokens MiMo-V2.6-Flash (Xiaomi): overall 37.9, $0.28 per million output tokens GPT-5.6 Luna (OpenAI): overall 37.3, $1.20 per million output tokens DeepSeek V4 Pro 0813 (DeepSeek): overall 36.0, $1.98 per million output tokens DeepSeek V4 Flash Vision Exp (DeepSeek): overall 34.8, $0.6468 per million output tokens DeepSeek V4 Flash 0731 (DeepSeek): overall 34.3, $1.32 per million output tokens Gemini 3.6 Flash (Google): overall 34.0, $3.75 per million output tokens Qwen3.8 27B (Qwen): overall 33.7, $2.20 per million output tokens GLM 5.2 (Z.ai): overall 33.7, $2.42 per million output tokens Muse Spark 1.1 (Meta): overall 33.7, $4.25 per million output tokens Gemini 3.5 Flash (Google): overall 33.6, $9.00 per million output tokens GPT-5.3-Codex (OpenAI): overall 32.5, $14.00 per million output tokens Claude Opus 4.6 (Anthropic): overall 31.9, $25.00 per million output tokens GPT-5.2 (OpenAI): overall 30.4, $14.00 per million output tokens DeepSeek V4 Pro 0423 (DeepSeek): overall 30.4, $1.9105 per million output tokens Claude Sonnet 4.6 (Anthropic): overall 30.1, $15.00 per million output tokens Gemini 3.1 Pro Preview (Google): overall 29.7, $12.00 per million output tokens Qwen3.7 Max (Qwen): overall 29.5, $4.425 per million output tokens MiniMax M3 (MiniMax): overall 29.2, $1.20 per million output tokens Claude Opus 4.5 (Anthropic): overall 29.1, $25.00 per million output tokens GPT-5.2-Codex (OpenAI): overall 28.5, $14.00 per million output tokens Solar Pro 4 (Upstage): overall 28.2, $0.60 per million output tokens GLM 5 (Z.ai): overall 27.9, $1.92 per million output tokens Grok Build 0.1 (xAI): overall 27.2, $2.00 per million output tokens Kimi K2.6 (Moonshot AI): overall 27.0, $3.41 per million output tokens Qwen3.6 Plus (Qwen): overall 27.0, $1.95 per million output tokens GLM 5 Turbo (Z.ai): overall 26.6, $4.00 per million output tokens Nemotron 3.5 Super VL (NVIDIA): overall 26.5, $0.50 per million output tokens Gemini 3 Flash Preview (Google): overall 26.3, $3.00 per million output tokens GLM 5.1 (Z.ai): overall 26.1, $3.036 per million output tokens MiMo-V2.5-Pro (Xiaomi): overall 26.0, $0.87 per million output tokens Kimi K2.7 Code (Moonshot AI): overall 25.8, $3.35 per million output tokens Inkling Small (Thinking Machines): overall 25.7, $1.20 per million output tokens Grok 4.20 (xAI): overall 25.7, $2.50 per million output tokens Hy3 preview (Tencent): overall 25.3, $0.60 per million output tokens Qwen3.7 Plus (Qwen): overall 25.2, $1.28 per million output tokens Inkling (Thinking Machines): overall 25.0, $4.05 per million output tokens Grok 4.3 (xAI): overall 24.9, $2.50 per million output tokens DeepSeek V4.1 Flash (DeepSeek): overall 24.7, $0.0112 per million output tokens GPT-5.1 (OpenAI): overall 24.7, $10.00 per million output tokens Ling 3.0 Flash VL (inclusionAI): overall 24.6, $0.18 per million output tokens DeepSeek V4 Flash 0423 (DeepSeek): overall 24.4, $0.182 per million output tokens GPT-5.4 Mini (OpenAI): overall 24.1, $4.50 per million output tokens Solar Mini 4 (Upstage): overall 24.1, $0.20 per million output tokens GPT-5.1-Codex (OpenAI): overall 23.7, $10.00 per million output tokens Kimi K2.5 (Moonshot AI): overall 23.5, $2.50 per million output tokens GLM 5V Turbo (Z.ai): overall 23.5, $4.00 per million output tokens GPT-5 (OpenAI): overall 23.0, $10.00 per million output tokens Qwen3.5-27B (Qwen): overall 22.9, $2.16 per million output tokens Nemotron 3 Ultra (NVIDIA): overall 22.9, $2.40 per million output tokens MiniMax M2.5 (MiniMax): overall 22.8, $0.95 per million output tokens MiniMax M2.7 (MiniMax): overall 22.8, $1.20 per million output tokens Claude Opus 4.1 (Anthropic): overall 22.8, $75.00 per million output tokens Gemini 3.5 Flash Lite (Google): overall 22.2, $2.50 per million output tokens GLM 4.7 (Z.ai): overall 22.2, $1.98 per million output tokens Kimi K2 Thinking (Moonshot AI): overall 22.0, $2.50 per million output tokens o3 Pro (OpenAI): overall 21.9, $80.00 per million output tokens MiMo-V2.5 (Xiaomi): overall 21.8, $0.28 per million output tokens DeepSeek V3.2 (DeepSeek): overall 21.5, $0.42 per million output tokens Qwen3.5 397B A17B (Qwen): overall 21.4, $2.34 per million output tokens Qwen3.6 27B (Qwen): overall 21.4, $2.00 per million output tokens MiniMax M2.1 (MiniMax): overall 20.9, $1.20 per million output tokens Claude Sonnet 4.5 (Anthropic): overall 20.7, $15.00 per million output tokens GPT-5.4 Nano (OpenAI): overall 20.7, $1.25 per million output tokens GPT-5 Mini (OpenAI): overall 20.6, $2.00 per million output tokens GPT-5.1-Codex-Mini (OpenAI): overall 20.4, $2.00 per million output tokens o3 (OpenAI): overall 20.2, $8.00 per million output tokens Ling 3.0 Flash (inclusionAI): overall 20.1, $0.063 per million output tokens Step 3.7 Flash (StepFun): overall 19.5, $1.15 per million output tokens Qwen3.5-35B-A3B (Qwen): overall 19.3, $1.00 per million output tokens LongCat 2.0 (Meituan): overall 19.1, $1.20 per million output tokens MiniMax M2 (MiniMax): overall 18.6, $1.02 per million output tokens GLM 4.6 (Z.ai): overall 18.5, $2.20 per million output tokens Qwen3.6 35B A3B (Qwen): overall 18.2, $0.95 per million output tokens Qwen3.5-122B-A10B (Qwen): overall 17.7, $2.08 per million output tokens Muse Glimmer 30B (Meta): overall 17.5, $1.20 per million output tokens Step 3.5 Flash (StepFun): overall 17.0, $0.30 per million output tokens Claude Haiku 4.5 (Anthropic): overall 16.9, $5.00 per million output tokens o4 Mini (OpenAI): overall 16.7, $4.40 per million output tokens Gemma 4 26B A4B (Google): overall 16.7, $0.255 per million output tokens DeepSeek V3.2 Exp (DeepSeek): overall 16.6, $0.41 per million output tokens Gemini 2.5 Pro (Google): overall 16.1, $10.00 per million output tokens Gemini 3.1 Flash Lite Preview (Google): overall 15.6, $1.50 per million output tokens Gemini 2.5 Flash (Google): overall 15.5, $2.50 per million output tokens Kimi K2 0905 (Moonshot AI): overall 15.3, $2.50 per million output tokens o1 (OpenAI): overall 15.2, $60.00 per million output tokens GLM 4.7 Flash (Z.ai): overall 14.9, $0.40 per million output tokens DeepSeek V3.1 Terminus (DeepSeek): overall 14.8, $1.00 per million output tokens Gemma 4 31B (Google): overall 14.7, $0.40 per million output tokens Mistral Medium 3.5 (Mistral AI): overall 14.2, $7.50 per million output tokens DeepSeek V3.1 (DeepSeek): overall 13.7, $0.95 per million output tokens Nova 2 Lite (Amazon): overall 13.4, $2.50 per million output tokens Qwen3.5-9B (Qwen): overall 13.3, $0.15 per million output tokens R1 0528 (DeepSeek): overall 13.1, $2.15 per million output tokens Command A (Cohere): overall 13.1, $10.00 per million output tokens GPT-5 Nano (OpenAI): overall 13.0, $0.40 per million output tokens Nemotron 3.5 Lightning (NVIDIA): overall 12.9, $0.20 per million output tokens GLM 4.5 (Z.ai): overall 12.8, $2.20 per million output tokens Nemotron 3 Super (NVIDIA): overall 12.8, $0.45 per million output tokens Kimi K2 0711 (Moonshot AI): overall 12.7, $2.30 per million output tokens GPT-4.1 (OpenAI): overall 12.7, $8.00 per million output tokens Qwen3 235B A22B Thinking 2507 (Qwen): overall 12.7, $3.50 per million output tokens o3 Mini (OpenAI): overall 12.5, $4.40 per million output tokens o1-pro (OpenAI): overall 12.4, $600.00 per million output tokens Mercury 2.5 (Inception): overall 12.3, $0.15 per million output tokens Qwen3 235B A22B Instruct 2507 (Qwen): overall 12.0, $0.55 per million output tokens Qwen3 Coder 480B A35B (Qwen): overall 11.9, $1.00 per million output tokens Sonar Reasoning Pro (Perplexity): overall 11.8, $8.00 per million output tokens MiniMax M1 (MiniMax): overall 11.7, $2.20 per million output tokens gpt-oss-120b (OpenAI): overall 11.6, $0.60 per million output tokens Mistral Small 4 (Mistral AI): overall 11.3, $0.60 per million output tokens Qwen3 Next 80B A3B Thinking (Qwen): overall 11.2, $1.20 per million output tokens GLM 4.6V (Z.ai): overall 11.2, $0.90 per million output tokens Granite 4.2 8B (IBM Granite): overall 11.1, $0.25 per million output tokens GLM 4.5 Air (Z.ai): overall 11.1, $0.85 per million output tokens o3 Mini High (OpenAI): overall 11.0, $4.40 per million output tokens Trinity Large Thinking (Arcee AI): overall 10.8, $0.80 per million output tokens Mercury 2 (Inception): overall 10.8, $0.75 per million output tokens Gemini 2.5 Flash Lite (Google): overall 10.4, $0.40 per million output tokens GPT-4.1 Mini (OpenAI): overall 10.2, $1.60 per million output tokens gpt-oss-20b (OpenAI): overall 10.0, $0.09 per million output tokens Qwen3 VL 235B A22B Instruct (Qwen): overall 9.9, $1.90 per million output tokens DeepSeek V3 0324 (DeepSeek): overall 9.7, $1.14 per million output tokens Qwen3 Coder 30B A3B Instruct (Qwen): overall 9.6, $0.28 per million output tokens Qwen3 Next 80B A3B Instruct (Qwen): overall 9.6, $1.10 per million output tokens Qwen3 VL 30B A3B Thinking (Qwen): overall 9.5, $1.00 per million output tokens Llama 4 Maverick (Meta Llama): overall 9.3, $0.6525 per million output tokens Mistral Large 3 2512 (Mistral AI): overall 9.3, $1.50 per million output tokens Nova Premier 1.0 (Amazon): overall 9.2, $12.50 per million output tokens Mistral Medium 3.1 (Mistral AI): overall 9.2, $2.00 per million output tokens Qwen3 Coder Next (Qwen): overall 9.2, $0.80 per million output tokens Mistral Medium 3 (Mistral AI): overall 9.0, $2.00 per million output tokens Nemotron 3 Nano 30B A3B (NVIDIA): overall 8.9, $0.24 per million output tokens Devstral 2 2512 (Mistral AI): overall 8.6, $2.00 per million output tokens GPT-4o (2024-11-20) (OpenAI): overall 8.4, $10.00 per million output tokens Qwen3 VL 30B A3B Instruct (Qwen): overall 7.9, $0.60 per million output tokens GPT-4.1 Nano (OpenAI): overall 7.8, $0.40 per million output tokens Solar Pro 3 (Upstage): overall 7.8, $0.60 per million output tokens GPT-4o (2024-08-06) (OpenAI): overall 7.7, $10.00 per million output tokens Llama 3.3 70B Instruct (Meta Llama): overall 7.7, $0.32 per million output tokens Sonar (Perplexity): overall 7.7, $1.00 per million output tokens Qwen2.5 72B Instruct (Qwen): overall 7.7, $0.40 per million output tokens Mistral Large 2407 (Mistral AI): overall 7.6, $6.00 per million output tokens Sonar Pro (Perplexity): overall 7.6, $15.00 per million output tokens Qwen3 30B A3B (Qwen): overall 7.6, $0.50 per million output tokens GLM 4.5V (Z.ai): overall 7.6, $1.80 per million output tokens Hermes 4 405B (Nous Research): overall 7.5, $3.00 per million output tokens Qwen3 30B A3B Instruct 2507 (Qwen): overall 7.5, $0.30 per million output tokens Qwen3 32B (Qwen): overall 7.3, $0.28 per million output tokens GPT-4o (2024-05-13) (OpenAI): overall 7.3, $15.00 per million output tokens Qwen3 VL 8B Instruct (Qwen): overall 7.3, $0.75 per million output tokens Nova Pro 1.0 (Amazon): overall 7.0, $3.20 per million output tokens GPT-4 Turbo (OpenAI): overall 7.0, $30.00 per million output tokens Llama 3.1 8B Instruct (Meta Llama): overall 6.9, $0.08 per million output tokens Qwen2.5 Coder 32B Instruct (Qwen): overall 6.7, $1.00 per million output tokens GPT-4 (OpenAI): overall 6.7, $60.00 per million output tokens Nova Lite 1.0 (Amazon): overall 6.7, $0.24 per million output tokens Mistral Small 3 (Mistral AI): overall 6.7, $0.08 per million output tokens Hermes 3 70B Instruct (Nous Research): overall 6.7, $0.70 per million output tokens GPT-4o-mini (OpenAI): overall 6.7, $0.60 per million output tokens Qwen3 14B (Qwen): overall 6.7, $0.24 per million output tokens Llama 3.1 70B Instruct (Meta Llama): overall 6.6, $0.40 per million output tokens Llama 4 Scout (Meta Llama): overall 6.5, $0.30 per million output tokens Saba (Mistral AI): overall 6.5, $0.60 per million output tokens Reka Flash 3 (Reka): overall 6.4, $0.20 per million output tokens Ministral 3 14B 2512 (Mistral AI): overall 6.0, $0.20 per million output tokens Nova Micro 1.0 (Amazon): overall 5.9, $0.14 per million output tokens Phi 4 (Microsoft): overall 5.9, $0.14 per million output tokens Mistral Large (Mistral AI): overall 5.8, $6.00 per million output tokens Llama 3.2 3B Instruct (Meta Llama): overall 5.7, $0.335 per million output tokens Mixtral 8x22B Instruct (Mistral AI): overall 5.7, $6.00 per million output tokens Ministral 3 8B 2512 (Mistral AI): overall 5.5, $0.15 per million output tokens GPT-3.5 Turbo (OpenAI): overall 5.5, $1.50 per million output tokens Granite 4.0 Micro (IBM Granite): overall 5.1, $0.112 per million output tokens Gemma 3 27B (Google): overall 4.9, $0.35 per million output tokens Gemma 3 4B (Google): overall 4.8, $0.05 per million output tokens Llama 3.2 1B Instruct (Meta Llama): overall 4.8, $0.201 per million output tokens Ministral 3 3B 2512 (Mistral AI): overall 4.8, $0.10 per million output tokens Gemma 3 12B (Google): overall 3.8, $0.15 per million output tokens Claude Opus 5.5Claude Sonnet 5.5Muse Spark 1.3MiMo-V2.6-ProClaude Haiku 5.5GLM 5.3 Flash Output price per million tokens, log scale Overall index
  • OpenAI
  • Qwen
  • Google
  • Mistral AI
  • DeepSeek
  • Anthropic
  • MiniMax
  • Meta, Meta Llama
  • Tencent
  • NVIDIA
  • Xiaomi
  • Perplexity
  • Cohere
  • Amazon
  • Upstage
  • StepFun
  • inclusionAI
  • Microsoft
  • IBM Granite
  • Arcee AI
  • Meituan
  • xAI, Z.ai, Moonshot AI, Inception, Nous Research, Reka
  • Other makers
213 paid models. A dot below the line has a model at least as good for the same money or less. Hover a dot for its numbers; click it for the model page. Overall: Artificial Analysis Intelligence Index. Price: OpenRouter.

The top two, category by category

The two highest models overall. Every pair of ranked models has a page like this; build one or open a popular one.

Claude Opus 5.5Anthropic · $20.00 out3 VS 1Claude Sonnet 5.5Anthropic · $10.00 out
  • 57.6 Overall 56.0
  • 92.2 Coding 88.9
  • No data Agentic No data
  • 90.6% Reasoning 95.6%
  • 100.0% Math 100.0%
  • No data Knowledge No data
  • 1336 Web design 1326

The big number is categories won, counted only where both have a result.Full comparison

Build your own

Search every model

What a working answer costs

Our own metric, published nowhere else: one SWE-bench Verified attempt divided by the share that succeed. Known for 84 models. How it is computed

The working price, model by model

Cost per solved SWE-bench Verified task. Lower is better.

1 of 84, cheapest first DeepSeek V4 Flash 0731 $0.01 per solved task DeepSeek · resolves 88.8%
The 84 models with a cost per solved task, around the ring. Source: OpenRouter, SWE-bench (swebench.com), and Vals AI boards, fetched October 20, 2026.

Why a cheap token is not a cheap answer

Cost per solved task is our own number: what one SWE-bench Verified attempt costs, divided by the share of attempts that succeed. GPT-4.1 Mini costs 5.0x less per token than GPT-4.1 and 4.9x more per solved task.

The five cheapest working answers

  1. 1 DeepSeek V4 Flash 0731 resolves 88.8% $0.01
  2. 2 MiMo-V2.5 resolves 71.0% $0.01
  3. 3 GLM 5.3 Flash resolves 92.0% $0.02
  4. 4 MiMo-V2.5-Pro resolves 74.0% $0.03
  5. 5 GPT-5.6 Luna resolves 93.0% $0.04

See every model on the cost and quality chart

Top five in every category

One row per public board, shown as published: the five best at their own results, the leader with its lead over second. All leaderboards · Best for a job

  1. CodingVibe Coding Index · 136 measured

    1. 1 Claude Opus 5.5Leads by 3.1 points 92.2
    2. 2 Claude Fable 5.1 89.1
    3. 3 Claude Sonnet 5.5 88.9
    4. 4 GPT-6.1 Sol 88.7
    5. 5 GPT-6 Astra 88.7

  2. AgenticAA Agentic Index · 87 measured

    1. 1 Claude Fable 5.1Leads by 1.4 points 57.9
    2. 2 Claude Opus 5 56.5
    3. 3 Qwen3.8 Max (0902) 56.0
    4. 4 Muse Spark 1.3 55.5
    5. 5 GLM 5.3 53.1

  3. ReasoningGPQA Diamond · 203 measured

    1. 1 Gemini 4 ArgonLeads by 0.6 points 96.4%
    2. 2 GPT-6 Astra 95.8%
    3. 3 Claude Sonnet 5.5 95.6%
    4. 4 Gemini 3.1 Pro Preview 95.5%
    5. 5 GPT-6.1 Sol 95.4%

  4. MathOTIS Mock AIME · 116 measured

    1. 1 Gemini 4 ArgonTied at the top with 9 more 100.0%
    2. 2 GPT-6.1 Sol 100.0%
    3. 3 Claude Sonnet 5.5 100.0%
    4. 4 GPT-6 Sol 100.0%
    5. 5 Claude Opus 5.5 100.0%

  5. KnowledgeMMLU Pro · 147 measured

    1. 1 Claude Fable 5.1Leads by 0.8 points 92.4%
    2. 2 Claude Opus 5 91.6%
    3. 3 Claude Fable 5 91.5%
    4. 4 Gemini 3.1 Pro Preview 91.0%
    5. 5 Gemini 3.8 Flash 90.2%

  6. Web designDesign Arena: websites · 131 measured

    1. 1 Muse WonderLeads by 4 Elo 1360
    2. 2 Muse Spark 1.3 1356
    3. 3 GPT-6 Astra 1338
    4. 4 Claude Opus 5.5 1336
    5. 5 Kimi K3 1335

The makers' medal table

In every category the makers are placed by their best model: gold, silver, bronze. Equal scores share a place, and a maker with no model on a board says so.

Makers placed by their best model in each category, with gold, silver and bronze counts
Maker 1Gold2Silver3Bronze OverallCodingAgenticReasoningMathKnowledgeWeb design
Anthropic 502 157.6 Claude Opus 5.5 192.2 Claude Opus 5.5 157.9 Claude Fable 5.1 395.6% Claude Sonnet 5.5 1100.0% Claude Sonnet 5.5 192.4% Claude Fable 5.1 31336 Claude Opus 5.5
Google 211 352.6 Gemini 4 Argon 583.1 Gemini 2.5 Pro Preview 06-05 940.2 Gemini 3.8 Flash 196.4% Gemini 4 Argon 1100.0% Gemini 4 Argon 291.0% Gemini 3.1 Pro Preview 51315 Gemini 4 Argon
OpenAI 140 252.7 GPT-6 Astra 288.7 GPT-6.1 Sol 651.0 GPT-6 Astra 295.8% GPT-6 Astra 1100.0% GPT-6.1 Sol 589.1% GPT-5.6 Sol 21338 GPT-6 Astra
Qwen 111 745.4 Qwen3.8 Max (0902) 387.4 Qwen3.8 Flash 256.0 Qwen3.8 Max (0902) 693.7% Qwen3.8 Max (0902) 1100.0% Qwen3.8 Max (0902) 489.3% Qwen3.7 Max 101276 Qwen3.8 Max (0902)
Meta 101 448.1 Muse Spark 1.3 679.7 Muse Spark 1.3 355.5 Muse Spark 1.3 594.1% Muse Spark 1.3 599.2% Muse Spark 1.3 688.7% Muse Spark 1.1 11360 Muse Wonder
xAI 001 546.4 Grok 4.7 776.4 Grok 4.6 553.0 Grok 4.6 494.7% Grok 4.6 599.2% Grok 4.6 389.4% Grok 4.6 91289 Grok 4.6
Xiaomi 000 646.3 MiMo-V2.6-Pro 972.6 MiMo-V2.6-Pro 1421.3 MiMo-V2.5-Pro 1882.6% MiMo-V2.5-Pro Not measured 1384.6% MiMo-V2.5-Pro 71308 MiMo-V2.6-Pro
Z.ai 000 844.8 GLM 5.3 873.1 GLM 5.3 453.1 GLM 5.3 1288.1% GLM 5.3 1093.9% GLM 5.3 Flash 1086.9% GLM 5.1 81295 GLM 5.3
StepFun 000 943.7 Step 5 Preview 1071.4 Step 5 Preview Not measured 1783.1% Step 3.5 Flash Not measuredNot measured 151196 Step 3.7 Flash
Moonshot AI 000 1043.6 Kimi K3 1264.7 Kimi K2 0711 750.0 Kimi K3 792.9% Kimi K3 997.2% Kimi K3 788.0% Kimi K3 41335 Kimi K3
inclusionAI 000 1141.1 Ling 3.1 Flash Not measured 1128.7 Ling 3.0 Flash VL 1586.2% Ling 3.0 Flash VL Not measured 1482.0% Ling 3.0 Flash Not measured
DeepSeek 000 1239.5 DeepSeek V4.1 Flash 1017 483.3 DeepSeek V4.1 Flash 1017 847.5 DeepSeek V4 Flash Vision Exp 992.4% DeepSeek V4 Pro 0813 798.6% DeepSeek V4 Pro 0813 887.3% DeepSeek V4 Pro 0423 61312 DeepSeek V4.1 Flash 1017

Best model for your job

The top three for each kind of work, each from the board named on the card. Every best-for list

Today's records

The fastest, the newest, the longest memory, the cheapest working answer, the biggest maker, the hottest download and the best open model, each with the board it was read from. All new and trending models

The state of play, October 2026

points between the top two
1.6
points cover the top ten
10.0
makers in the top ten
4

Anthropic Claude Opus 5.5 leads the overall table with 57.6 on the Artificial Analysis Intelligence Index, 1.6 points ahead of Anthropic Claude Sonnet 5.5. The top ten are within 10.0 points of each other and come from 4 makers: Anthropic, OpenAI, Google and Meta.

Price does not follow rank. Inside the top ten, Muse Spark 1.3 costs $4.25 per million output tokens and scores 48.1, while Claude Fable 5.1 costs $50.00. For coding work, the cost per solved task tells you more than the token price, because a cheap model that fails costs you the retry.

The best model with public weights, MiMo-V2.6-Pro, sits at place 13 of 213 with 46.3. Coding has its own leader: Claude Opus 5.5 tops our Vibe Coding Index, which averages four public coding benchmarks and shows how many of them stand behind each score.

43 models were released in the last 30 days, the newest DeepSeek V4.1 Flash 1017 on October 17, 2026. New models often appear here before every board has measured them, so their rows show "No data" until the results are published.

Sources, read today

Every board we read, most models first: how many models it gave us, the bar against the biggest, and when the board itself last changed. Methodology

  1. 356OpenRouter read October 20, 2026
  2. 213Artificial Analysis (via OpenRouter) read October 20, 2026
  3. 198Epoch AI benchmarking hub read October 20, 2026
  4. 195Artificial Analysis read October 20, 2026
  5. 175OpenRouter (provider speed) read October 20, 2026
  6. 169LMArena (Hugging Face dataset) read October 20, 2026
  7. 151Hugging Face read October 20, 2026
  8. 141Design Arena read October 20, 2026
  9. 132Vals AI boards read October 20, 2026
  10. 104ALE-Bench (Sakana AI) read October 20, 2026
  11. 101LMArena WebDev read October 20, 2026
  12. 92MMLU Pro (Vals AI) read October 20, 2026, board updated September 1, 2026
  13. 91GPQA (Vals AI) read October 20, 2026, board updated September 1, 2026
  14. 54AIME (Vals AI) read October 20, 2026, board updated April 16, 2026
  15. 43LiveBench read October 20, 2026, board updated October 14, 2026
  16. 38SWE-bench (swebench.com) read October 20, 2026
  17. 22Aider polyglot leaderboard read October 20, 2026
  18. 20Vals Index read October 20, 2026, board updated October 19, 2026

Questions

What is the best LLM right now?

As of October 20, 2026, Anthropic Claude Opus 5.5 leads 213 models overall with 57.6 on the Artificial Analysis Intelligence Index. The best model depends on the job, so the tabs above re-rank the table by coding, agents, reasoning, math, knowledge and web design.

Which LLM is best for coding?

Anthropic Claude Opus 5.5 leads our Vibe Coding Index with 92.2, from two benchmarks. 136 models have a coding score; the dots on each row show how many public coding benchmarks stand behind it.

Which LLM is best for agents and tool use?

Anthropic Claude Fable 5.1 leads the Artificial Analysis Agentic Index with 57.9 as of October 20, 2026. The index covers tasks where a model works over several steps and calls tools.

What is the best cheap LLM?

Among models at or under $1 per million output tokens on OpenRouter, Xiaomi MiMo-V2.6-Pro has the highest overall index, 46.3, at $0.87. For coding, the cost per solved task is the better measure of cheap.

What is the best open-weight LLM?

Xiaomi MiMo-V2.6-Pro has the highest overall index among models with public weights on Hugging Face, 46.3. We link the repository; read its license there.

Where do the numbers come from, and how often are they updated?

From 18 public sources, read every day: OpenRouter for the catalog and prices, the public leaderboards for benchmark results. Every number on the site names its source and the day it was read. We do not run benchmarks ourselves.

Why do some cells say "No data"?

Because that board has not measured that model. We leave the gap visible instead of estimating a number, so every figure you see was published by someone you can check.

T T

Every badge is a model

The top 120 models on the overall index spell the name, number one at the top of the first L. Hover a badge for the model and its place, click it for its page.