AI Indexer

USD per 1M tokens, blended 3:1
Artificial Analysis Intelligence Index
723 models

Intelligence

Highest Intelligence Index · higher is better

Artificial Analysis
Gemini 4 Carbon (High): 5959Gemini 4 Carb…Claude Opus 5.5 (Max, Default Fallback): 5858Claude Opus 5…Gemini 4 Carbon (Medium): 5858Gemini 4 Carb…Claude Opus 5.5 (Xhigh, Default Fallback): 5757Claude Opus 5…GPT-6 Astra (Max): 5757GPT-6 Astra (…GPT-6 Astra (Xhigh): 5555GPT-6 Astra (…GPT-6 Astra (High): 5555GPT-6 Astra (…GPT-6.1 Sol (Max): 5555GPT-6.1 Sol (…Claude Sonnet 5.5 (Max, Default Fallback): 5454Claude Sonnet…Gemini 4 Carbon (Low): 5454Gemini 4 Carb…GPT-6.1 Sol (Xhigh): 5252GPT-6.1 Sol (…Claude Opus 5.5 (High, Default Fallback): 5252Claude Opus 5…
  • Google
  • Anthropic
  • OpenAI

Speed

Fastest output among capable models (index 30+) · tokens per second

Artificial Analysis
Claude Haiku 5.5 (Max): 229229Claude Haiku …Claude Haiku 5.5 (Xhigh): 174174Claude Haiku …DeepSeek V4.1 Flash 1017 (Max): 156156DeepSeek V4.1…Muse Wonder (Max): 139139Muse Wonder (…Claude Sonnet 5.5 (Max, Default Fallback): 133133Claude Sonnet…Gemini 3.8 Flash (High): 130130Gemini 3.8 Fl…GPT-6.1 Terra (Max): 117117GPT-6.1 Terra…Claude Opus 5.5 (Max, Default Fallback): 9696Claude Opus 5…GPT-6.1 Terra (Xhigh): 9595GPT-6.1 Terra…Claude Sonnet 5.5 (Xhigh, Default Fallback): 9595Claude Sonnet…GPT-6.1 Terra (Medium): 9191GPT-6.1 Terra…Claude Sonnet 5.5 (Medium, Default Fallback): 9191Claude Sonnet…
  • Anthropic
  • DeepSeek
  • Meta
  • Google
  • OpenAI

Cost

Cheapest capable models (index 35+) · USD per 1M tokens, blended · lower is better

Artificial Analysis
Claude Haiku 5.5 (Max): $0.2$0.2Claude Haiku …GPT-6.1 Terra (Max): $2.00$2.00GPT-6.1 Terra…GPT-6.1 Terra (Xhigh): $2.00$2.00GPT-6.1 Terra…GPT-6.1 Terra (High): $2.00$2.00GPT-6.1 Terra…GPT-6.1 Terra (Medium): $2.00$2.00GPT-6.1 Terra…Muse Spark 1.3 (Max): $2.00$2.00Muse Spark 1.…Grok 4.7 (Xhigh): $3.00$3.00Grok 4.7 (Xhi…Grok 4.7 (High): $3.00$3.00Grok 4.7 (Hig…GPT-6.1 Sol (Max): $4.00$4.00GPT-6.1 Sol (…Claude Sonnet 5.5 (Max, Default Fallback): $4.00$4.00Claude Sonnet…GPT-6.1 Sol (Xhigh): $4.00$4.00GPT-6.1 Sol (…GPT-6.1 Sol (High): $4.00$4.00GPT-6.1 Sol (…
  • Anthropic
  • OpenAI
  • Meta
  • xAI

Best offline models

Open weights: download them and run them on your own computer or server · Intelligence Index

Artificial Analysis
Kimi K3.1 Pro: 4646Kimi K3.1 ProDeepSeek V4.1 Flash 1017 (Max): 3333DeepSeek V4.1…GLM-5.3 (Max): 3232GLM-5.3 (Max)MiMo-V2.6-Pro: 3030MiMo-V2.6-ProMiniMax-M3.1-Flash: 2929MiniMax-M3.1-…DeepSeek V4.1 Flash 1017 (High): 2929DeepSeek V4.1…Step 5 Preview: 2929Step 5 PreviewGLM-5.3-Flash: 2828GLM-5.3-FlashMiMo-V2.6-Flash: 2727MiMo-V2.6-Fla…Mistral Large 4 (High): 2626Mistral Large…Mistral Large 4 Preview: 2525Mistral Large…Muse Spark Open (Xhigh): 2525Muse Spark Op…
  • Kimi
  • DeepSeek
  • Z.ai
  • Xiaomi
  • MiniMax
  • Stepfun
  • Mistral
  • Meta
  1. 1

    Gemini 4 Carbon

    High
    Google Reasoning
    Intelligence
    59.4
    Coding
    —
    Speed
    91 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 63.06s
  2. 2

    Claude Opus 5.5

    Max, Default Fallback
    Anthropic Reasoning
    Intelligence
    58.4
    Coding
    —
    Speed
    96 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 565.76sCtx 1M
  3. 3

    Gemini 4 Carbon

    Medium
    Google Reasoning
    Intelligence
    57.7
    Coding
    —
    Speed
    68 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 10.50s
  4. 4

    Claude Opus 5.5

    Xhigh, Default Fallback
    Anthropic Reasoning
    Intelligence
    57.0
    Coding
    —
    Speed
    78 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 71.43sCtx 1M
  5. 5

    GPT-6 Astra

    Max
    OpenAI Reasoning
    Intelligence
    56.9
    Coding
    —
    Speed
    59 t/s
    Price
    $20/1M
    In $10Out $50TTFT 213.14sCtx 1.05M
  6. 6

    GPT-6 Astra

    Xhigh
    OpenAI Reasoning
    Intelligence
    55.3
    Coding
    —
    Speed
    55 t/s
    Price
    $20/1M
    In $10Out $50TTFT 89.29sCtx 1.05M
  7. 7

    GPT-6 Astra

    High
    OpenAI Reasoning
    Intelligence
    54.8
    Coding
    —
    Speed
    51 t/s
    Price
    $20/1M
    In $10Out $50TTFT 19.27sCtx 1.05M
  8. 8

    GPT-6.1 Sol

    Max
    OpenAI Reasoning
    Intelligence
    54.5
    Coding
    —
    Speed
    68 t/s
    Price
    $4.00/1M
    In $2.00Out $10TTFT 166.61sCtx 1.05M
  9. 9

    Claude Sonnet 5.5

    Max, Default Fallback
    Anthropic Reasoning
    Intelligence
    54.2
    Coding
    —
    Speed
    133 t/s
    Price
    $4.00/1M
    In $2.00Out $10TTFT 309.42sCtx 1M
  10. 10

    Gemini 4 Carbon

    Low
    Google Reasoning
    Intelligence
    53.7
    Coding
    —
    Speed
    0 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 0.00s
  11. 11

    GPT-6.1 Sol

    Xhigh
    OpenAI Reasoning
    Intelligence
    52.3
    Coding
    —
    Speed
    60 t/s
    Price
    $4.00/1M
    In $2.00Out $10TTFT 47.70sCtx 1.05M
  12. 12

    Claude Opus 5.5

    High, Default Fallback
    Anthropic Reasoning
    Intelligence
    51.7
    Coding
    —
    Speed
    70 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 19.25sCtx 1M
  13. 13

    GPT-6 Astra

    Medium
    OpenAI Reasoning
    Intelligence
    51.5
    Coding
    —
    Speed
    48 t/s
    Price
    $20/1M
    In $10Out $50TTFT 3.25sCtx 1.05M
  14. 14

    GPT-6.1 Terra

    Max
    OpenAI Reasoning
    Intelligence
    51.0
    Coding
    —
    Speed
    117 t/s
    Price
    $2.00/1M
    In $1.00Out $5.00TTFT 240.73sCtx 1.05M
  15. 15

    Gemini 4 Argon

    High
    Google Reasoning
    Intelligence
    50.4
    Coding
    —
    Speed
    82 t/s
    Price
    $8.00/1M
    In $4.00Out $20TTFT 31.34sCtx 1.05M
  16. 16

    GPT-6.1 Sol

    High
    OpenAI Reasoning
    Intelligence
    50.2
    Coding
    —
    Speed
    57 t/s
    Price
    $4.00/1M
    In $2.00Out $10TTFT 17.40sCtx 1.05M
  17. 17

    Claude Fable 5.1

    Xhigh, Default Fallback
    Anthropic Reasoning
    Intelligence
    49.7
    Coding
    —
    Speed
    54 t/s
    Price
    $20/1M
    In $10Out $50TTFT 29.67sCtx 1M
  18. 18

    Muse Wonder Full

    Max
    Meta Reasoning
    Intelligence
    49.7
    Coding
    —
    Speed
    0 t/s
    Price
    Free
    In FreeOut FreeTTFT 0.00s
  19. 19

    Claude Sonnet 5.5

    Xhigh, Default Fallback
    Anthropic Reasoning
    Intelligence
    49.6
    Coding
    —
    Speed
    95 t/s
    Price
    $4.00/1M
    In $2.00Out $10TTFT 13.95sCtx 1M
  20. 20

    Claude Fable 5.1

    Max, Default Fallback
    Anthropic Reasoning
    Intelligence
    49.0
    Coding
    —
    Speed
    68 t/s
    Price
    $20/1M
    In $10Out $50TTFT 183.07sCtx 1M

Benchmarks, speed and prices: Artificial Analysis · capabilities and catalogue: OpenRouter

Output speed

Median output tokens per second · higher is better

Artificial Analysis
Claude Sonnet 5.5 (Max, Default Fallback): 133133Claude Sonnet…GPT-6.1 Terra (Max): 117117GPT-6.1 Terra…Claude Opus 5.5 (Max, Default Fallback): 9696Claude Opus 5…Claude Sonnet 5.5 (Xhigh, Default Fallback): 9595Claude Sonnet…Gemini 4 Carbon (High): 9191Gemini 4 Carb…Gemini 4 Argon (High): 8282Gemini 4 Argo…Claude Opus 5.5 (Xhigh, Default Fallback): 7878Claude Opus 5…Claude Opus 5.5 (Medium, Default Fallback): 7272Claude Opus 5…Claude Opus 5.5 (High, Default Fallback): 7070Claude Opus 5…Claude Fable 5.1 (Max, Default Fallback): 6868Claude Fable …Gemini 4 Carbon (Medium): 6868Gemini 4 Carb…GPT-6.1 Sol (Max): 6868GPT-6.1 Sol (…

Latency

Median seconds to first token · lower is better

Artificial Analysis
Gemini 4 Carbon (Low): 0.00s0.00sGemini 4 Carb…Muse Wonder Full (Max): 0.00s0.00sMuse Wonder F…Claude Fable 5.1 (High, Default Fallback): 2.53s2.53sClaude Fable …Kimi K3.1 Pro: 2.65s2.65sKimi K3.1 ProGPT-6 Astra (Medium): 3.25s3.25sGPT-6 Astra (…GPT-6.1 Sol (Medium): 3.65s3.65sGPT-6.1 Sol (…Claude Opus 5.5 (Medium, Default Fallback): 9.12s9.12sClaude Opus 5…Gemini 4 Carbon (Medium): 10.50s10.50sGemini 4 Carb…Claude Sonnet 5.5 (Xhigh, Default Fallback): 13.95s13.95sClaude Sonnet…GPT-6.1 Sol (High): 17.40s17.40sGPT-6.1 Sol (…Claude Opus 5.5 (High, Default Fallback): 19.25s19.25sClaude Opus 5…GPT-6 Astra (High): 19.27s19.27sGPT-6 Astra (…

Price

USD per 1M tokens for each lab's newest flagship-class model · input (left) and output (right) · lower is better

OpenRouter · live
Phi 4: $0.07$0.07Phi 4: $0.14$0.14Phi 4Claude Haiku 5.5: $0.1$0.1Claude Haiku 5.5: $0.5$0.5Claude Haiku …Nemotron 3.5 Super VL: $0.15$0.15Nemotron 3.5 Super VL: $0.5$0.5Nemotron 3.5 …Command A+: $0.3$0.3Command A+: $1.50$1.50Command A+Mistral Large 4: $0.7$0.7Mistral Large 4: $2.10$2.10Mistral Large…DeepSeek V4 Pro 0813: $0.12$0.12DeepSeek V4 Pro 0813: $4.00$4.00DeepSeek V4 P…Muse Spark Open: $1.25$1.25Muse Spark Open: $4.25$4.25Muse Spark Op…GPT-6.1 Terra: $1.00$1.00GPT-6.1 Terra: $5.00$5.00GPT-6.1 TerraGrok 4.7: $2.00$2.00Grok 4.7: $6.00$6.00Grok 4.7GLM 5.3 Prime: $2.80$2.80GLM 5.3 Prime: $8.80$8.80GLM 5.3 PrimeKimi K3.1 Pro: $2.00$2.00Kimi K3.1 Pro: $10$10Kimi K3.1 ProQwen3.8 Max Prime: $4.00$4.00Qwen3.8 Max Prime: $12$12Qwen3.8 Max P…
  • Microsoft
  • Anthropic
  • NVIDIA
  • Cohere
  • Mistral
  • DeepSeek
  • Meta
  • OpenAI
  • xAI
  • Z.ai
  • Moonshot
  • Alibaba

Context window

Tokens of input a model can read at once · larger is better

OpenRouter · live
Grok 4.20: 2M2MGrok 4.20Llama 4 Scout: 1.31M1.31MLlama 4 ScoutGPT-6.1 Terra: 1.05M1.05MGPT-6.1 TerraGLM 5.3 FlashX Legal: 1.05M1.05MGLM 5.3 Flash…Kimi K3.1 Pro: 1.05M1.05MKimi K3.1 ProMiniMax M3.1 Flash: 1.05M1.05MMiniMax M3.1 …DeepSeek V4.1 Flash 1017: 1.05M1.05MDeepSeek V4.1…Muse Spark Open: 1.05M1.05MMuse Spark Op…Nemotron 3.5 Super VL: 1.05M1.05MNemotron 3.5 …Gemini 4 Argon: 1.05M1.05MGemini 4 ArgonMistral Large 4: 1.05M1.05MMistral Large…Qwen3.8 2.4T A95B: 1.05M1.05MQwen3.8 2.4T …
  • xAI
  • Meta
  • OpenAI
  • Z.ai
  • Moonshot
  • MiniMax
  • DeepSeek
  • NVIDIA
  • Google
  • Mistral
  • Alibaba

Releases over time

One dot per model release in the last twelve months · most active labs

OpenRouter · live
DecJanFebMarAprMayJunJulAugSepOctNov OpenAI35 releases GPT-6.1 Terra · 26 Oct 2026GPT-6.1 Terra Pro · 26 Oct 2026GPT-6.1 Sol · 29 Sep 2026GPT-6.1 Sol Pro · 29 Sep 2026GPT-6 Luna · 22 Sep 2026GPT-6 Luna Pro · 22 Sep 2026GPT-6 Sol · 22 Sep 2026GPT-6 Sol Pro · 22 Sep 2026GPT-6 Astra · 4 Sep 2026GPT-6 Astra Pro · 4 Sep 2026GPT-5.6 Luna · 9 Jul 2026GPT-5.6 Luna Pro · 9 Jul 2026GPT-5.6 Sol · 9 Jul 2026GPT-5.6 Sol Pro · 9 Jul 2026GPT-5.6 Terra · 9 Jul 2026GPT-5.6 Terra Pro · 9 Jul 2026GPT Chat Latest · 5 May 2026GPT-5.5 · 24 Apr 2026GPT-5.5 Pro · 24 Apr 2026GPT-5.4 Image 2 · 21 Apr 2026GPT-5.4 Mini · 17 Mar 2026GPT-5.4 Nano · 17 Mar 2026GPT-5.4 · 5 Mar 2026GPT-5.4 Pro · 5 Mar 2026GPT-5.3-Codex · 24 Feb 2026GPT Audio · 19 Jan 2026GPT Audio Mini · 19 Jan 2026GPT-5.2-Codex · 14 Jan 2026GPT-5.2 · 10 Dec 2025GPT-5.2 Chat · 10 Dec 2025GPT-5.2 Pro · 10 Dec 2025GPT-5.1-Codex-Max · 4 Dec 2025GPT-5.1 · 13 Nov 2025GPT-5.1-Codex · 13 Nov 2025GPT-5.1-Codex-Mini · 13 Nov 2025 Alibaba22 releases Qwen3.8 Max Prime · 23 Sep 2026Qwen3.8 Omni Flash · 21 Sep 2026Qwen3.8 Max (0902) · 3 Sep 2026Qwen3.8 Flash · 26 Aug 2026Qwen3.8 27B · 14 Aug 2026Qwen3.8 2.4T A95B · 12 Aug 2026Qwen3.7 Flash · 27 Jul 2026Qwen3.7 Plus · 3 Jun 2026Qwen3.7 Max · 21 May 2026Qwen3.5 Plus 2026-04-20 · 27 Apr 2026Qwen3.6 27B · 27 Apr 2026Qwen3.6 35B A3B · 27 Apr 2026Qwen3.6 Flash · 27 Apr 2026Qwen3.6 Plus · 2 Apr 2026Qwen3.5-9B · 10 Mar 2026Qwen3.5-122B-A10B · 25 Feb 2026Qwen3.5-27B · 25 Feb 2026Qwen3.5-35B-A3B · 25 Feb 2026Qwen3.5-Flash · 25 Feb 2026Qwen3.5 397B A17B · 16 Feb 2026Qwen3.5 Plus 2026-02-15 · 16 Feb 2026Qwen3 Coder Next · 4 Feb 2026 Google21 releases Gemini 4 Argon · 13 Oct 2026Nano Banana 2.1 · 6 Oct 2026Gemini 3.8 Flash · 2 Sep 2026Gemini 3.7 Flash · 13 Aug 2026Gemini 3.5 Flash Lite · 21 Jul 2026Gemini 3.6 Flash · 21 Jul 2026Nano Banana 2 Lite (Gemini 3.1 Flash Lite Image) · 30 Jun 2026Nano Banana 2 (Gemini 3.1 Flash Image) · 18 Jun 2026Nano Banana Pro (Gemini 3 Pro Image) · 18 Jun 2026Gemini 3.5 Flash · 19 May 2026Gemini 3.1 Flash Lite · 7 May 2026Gemma 4 26B A4B · 3 Apr 2026Gemma 4 31B · 2 Apr 2026Lyria 3 Clip Preview · 30 Mar 2026Lyria 3 Pro Preview · 30 Mar 2026Gemini 3.1 Flash Lite Preview · 3 Mar 2026Nano Banana 2 (Gemini 3.1 Flash Image Preview) · 26 Feb 2026Gemini 3.1 Pro Preview Custom Tools · 25 Feb 2026Gemini 3.1 Pro Preview · 19 Feb 2026Gemini 3 Flash Preview · 17 Dec 2025Nano Banana Pro (Gemini 3 Pro Image Preview) · 20 Nov 2025 Z.ai13 releases GLM 5.3 FlashX Legal · 2 Nov 2026GLM 5.3 Prime · 23 Sep 2026GLM 5.3 FlashX · 18 Sep 2026GLM 5.3 Flash · 26 Aug 2026GLM 5.3 · 18 Aug 2026GLM 5.2 · 16 Jun 2026GLM 5.1 · 7 Apr 2026GLM 5V Turbo · 1 Apr 2026GLM 5 Turbo · 15 Mar 2026GLM 5 · 11 Feb 2026GLM 4.7 Flash · 19 Jan 2026GLM 4.7 · 22 Dec 2025GLM 4.6V · 8 Dec 2025 Anthropic12 releases Claude Haiku 5.5 · 7 Oct 2026Claude Sonnet 5.5 · 28 Sep 2026Claude Opus 5.5 · 22 Sep 2026Claude Fable 5.1 · 1 Sep 2026Claude Opus 5 · 24 Jul 2026Claude Sonnet 5 · 30 Jun 2026Claude Fable 5 · 9 Jun 2026Claude Opus 4.8 · 27 May 2026Claude Opus 4.7 · 16 Apr 2026Claude Sonnet 4.6 · 17 Feb 2026Claude Opus 4.6 · 4 Feb 2026Claude Opus 4.5 · 24 Nov 2025 DeepSeek8 releases DeepSeek V4.1 Flash 1017 · 17 Oct 2026DeepSeek V4.1 Flash · 10 Sep 2026DeepSeek V4 Flash Vision Exp · 21 Aug 2026DeepSeek V4 Pro 0813 · 12 Aug 2026DeepSeek V4 Flash 0731 · 31 Jul 2026DeepSeek V4 Flash 0423 · 24 Apr 2026DeepSeek V4 Pro 0423 · 24 Apr 2026DeepSeek V3.2 · 1 Dec 2025 Meta8 releases Muse Spark Open · 15 Oct 2026Muse Wonder · 12 Oct 2026Muse Spark 1.3 · 2 Sep 2026Muse Spark 1.3 Contributor · 2 Sep 2026Muse Spark 1.2 Contributor · 21 Aug 2026Muse Glimmer 30B · 9 Aug 2026Muse Spark 1.2 · 5 Aug 2026Muse Spark 1.1 · 16 Jul 2026 Mistral8 releases Mistral Large 4 · 6 Oct 2026Mistral Medium 3.5 · 30 Apr 2026Mistral Small 4 · 16 Mar 2026Devstral 2 2512 · 9 Dec 2025Ministral 3 14B 2512 · 2 Dec 2025Ministral 3 3B 2512 · 2 Dec 2025Ministral 3 8B 2512 · 2 Dec 2025Mistral Large 3 2512 · 1 Dec 2025 NVIDIA7 releases Nemotron 3.5 Super VL · 15 Oct 2026Switchyard · 21 Sep 2026Nemotron 3.5 Lightning · 11 Aug 2026Nemotron 3 Ultra · 4 Jun 2026Nemotron 3.5 Content Safety · 4 Jun 2026Nemotron 3 Super · 11 Mar 2026Nemotron 3 Nano 30B A3B · 14 Dec 2025 xAI7 releases Grok 4.7 · 21 Sep 2026Grok 4.6 · 12 Aug 2026Grok 4.5 · 8 Jul 2026Grok Build 0.1 · 20 May 2026Grok 4.3 · 30 Apr 2026Grok 4.20 · 31 Mar 2026Grok 4.20 Multi-Agent · 31 Mar 2026

Open weights vs. proprietary

Releases in the last twelve months per lab · open weights can be self-hosted

OpenRouter · live
  1. OpenAI0 open · 35 closed
  2. Alibaba11 open · 11 closed
  3. Google2 open · 19 closed
  4. Z.ai8 open · 5 closed
  5. Anthropic0 open · 12 closed
  6. DeepSeek8 open · 0 closed
  7. Meta2 open · 6 closed
  8. Mistral6 open · 2 closed
  9. NVIDIA6 open · 1 closed
  10. xAI0 open · 7 closed
44%of all 341 tracked models ship open weightsOpen weights Proprietary

Compare the frontier

Each lab's newest flagship-class model

OpenRouter · live
ModelLabReleasedContextInput $/1MOutput $/1MCapabilities
GPT-6.1 TerraOpenAI26 Oct 20261.05M$1.00$5.00
ReasoningVision
Claude Haiku 5.5Anthropic7 Oct 20261M$0.1$0.5
ReasoningVision
DeepSeek V4 Pro 0813DeepSeek12 Aug 20261.05M$0.12$4.00
ReasoningOpen weights
Qwen3.8 Max PrimeAlibaba23 Sep 20261M$4.00$12
ReasoningVisionVideo
Grok 4.7xAI21 Sep 2026500K$2.00$6.00
ReasoningVision
Muse Spark OpenMeta15 Oct 20261.05M$1.25$4.25
ReasoningVisionAudioVideoOpen weights
Mistral Large 4Mistral6 Oct 20261.05M$0.7$2.10
ReasoningVisionOpen weights
GLM 5.3 PrimeZ.ai23 Sep 20261M$2.80$8.80
Reasoning
Kimi K3.1 ProMoonshot23 Oct 20261.05M$2.00$10
ReasoningVisionVideo
Nova Premier 1.0Amazon31 Oct 20251M$2.50$13
Vision
Nemotron 3.5 Super VLNVIDIA15 Oct 20261.05M$0.15$0.5
ReasoningVisionVideoOpen weights
Phi 4Microsoft10 Jan 202516K$0.07$0.14
Open weights
Command A+Cohere22 Sep 2026192K$0.3$1.50
ReasoningVision