{ "name": "Kingy AI Price Index", "licence": "CC BY 4.0", "citation": "Kingy AI Price Index, May 2024 edition. Kingy AI. https://kingy.ai/ai-price-index/2024-05/", "url": "https://kingy.ai/ai-price-index/2024-05/", "method": { "version": "1.0", "url": "https://kingy.ai/ai-price-index/#method" }, "edition": { "edition": "2024-05", "label": "May 2024", "reconstructed": true, "referenceDate": "2024-05-01", "capturedAt": "2024-05-01T00:00:00+00:00", "method": "1.0", "previous": "2024-04", "previousReferenceDate": "2024-04-01", "intervalDays": 30, "headline": "In May 2024, a million AI tokens cost a median $15.00 from frontier models (no change) and $0.75 from small models (no change). Changes compare with April.", "story": { "title": "AI prices held steady in May", "text": "No tier of the index moved by 1% or more from April.", "tier": null }, "changes": { "likeForLike": [], "unchanged": 6, "basket": [ { "provider": "OpenAI", "tier": "frontier", "from": { "model": "GPT-4 Turbo", "value": "15.0000" }, "to": { "model": "GPT-4 Turbo", "value": "15.0000" } } ] }, "scheduled": [], "outlook": [], "withheld": [], "corrections": [], "tiers": [ { "id": "frontier", "label": "Frontier", "kind": "api", "summary": "Each lab's most capable model", "unit": "USD per 1M tokens, 3:1 input:output blend", "value": "15.0000", "previousValue": "15.0000", "changePct": 0, "n": 3, "min": "12.0000", "max": "30.0000", "constituents": [ "mistral-large", "openai-gpt-4-turbo", "anthropic-claude-3-opus" ], "published": true, "indexLevel": "333.33" }, { "id": "mid", "label": "Mid-range", "kind": "api", "summary": "Each lab's balanced workhorse model", "unit": "USD per 1M tokens, 3:1 input:output blend", "value": null, "previousValue": null, "changePct": null, "n": 1, "min": "6.0000", "max": "6.0000", "constituents": [ "anthropic-claude-3-sonnet" ], "published": false, "indexLevel": null }, { "id": "small", "label": "Small", "kind": "api", "summary": "Each lab's low-cost, high-volume model", "unit": "USD per 1M tokens, 3:1 input:output blend", "value": "0.7500", "previousValue": "0.7500", "changePct": 0, "n": 3, "min": "0.5000", "max": "3.0000", "constituents": [ "anthropic-claude-3-haiku", "openai-gpt-3-5-turbo", "mistral-small" ], "published": true, "indexLevel": "142.86" }, { "id": "standard", "label": "Standard plan", "kind": "subscription", "summary": "Each provider's main paid individual plan", "unit": "USD per month", "value": null, "previousValue": null, "changePct": null, "n": 0, "min": null, "max": null, "constituents": [], "published": false, "indexLevel": null }, { "id": "top", "label": "Top plan", "kind": "subscription", "summary": "Each provider's highest-priced individual plan", "unit": "USD per month", "value": null, "previousValue": null, "changePct": null, "n": 0, "min": null, "max": null, "constituents": [], "published": false, "indexLevel": null } ], "observations": [ { "id": "openai-gpt-4-turbo", "provider": "OpenAI", "model": "GPT-4 Turbo", "modelId": "gpt-4-turbo-2024-04-09", "tier": "frontier", "status": "verified", "input": "10.00", "output": "30.00", "blended": "15.0000", "source": { "url": "https://openai.com/pricing", "page": "https://openai.com/pricing", "fetchedAt": "2026-10-19T21:49:56+00:00", "snapshotUrl": "https://web.archive.org/web/20240427231923/https://openai.com/pricing", "snapshotAt": "2024-04-27T23:19:23+00:00", "sha256": "d006311dd64529a403e17d3a2eac8375757971232cfde7866cbfcc9a8640db4f", "excerpt": "gpt-4-turbo-2024-04-09 $10.00 / 1 M tokens $30.00" }, "positioning": { "url": "https://openai.com/pricing", "snapshotUrl": "https://web.archive.org/web/20240427231923/https://openai.com/pricing", "quote": "GPT-4 Turbo is more powerful than GPT-4", "rule": "OpenAI made no \"most capable\" claim on this page; GPT-4 Turbo is the newest model in its highest-priced line (GPT-4), and the page says it is more powerful than GPT-4." } }, { "id": "openai-gpt-3-5-turbo", "provider": "OpenAI", "model": "GPT-3.5 Turbo", "modelId": "gpt-3.5-turbo-0125", "tier": "small", "status": "verified", "input": "0.50", "output": "1.50", "blended": "0.7500", "source": { "url": "https://openai.com/pricing", "page": "https://openai.com/pricing", "fetchedAt": "2026-10-19T21:49:56+00:00", "snapshotUrl": "https://web.archive.org/web/20240427231923/https://openai.com/pricing", "snapshotAt": "2024-04-27T23:19:23+00:00", "sha256": "d006311dd64529a403e17d3a2eac8375757971232cfde7866cbfcc9a8640db4f", "excerpt": "gpt-3.5-turbo-0125 $0.50 / 1 M tokens $1.50" }, "positioning": { "url": "https://openai.com/pricing", "snapshotUrl": "https://web.archive.org/web/20240427231923/https://openai.com/pricing", "quote": "gpt-3.5-turbo-0125 is the flagship model of this family", "rule": null } }, { "id": "anthropic-claude-3-opus", "provider": "Anthropic", "model": "Claude 3 Opus", "modelId": "claude-3-opus-20240229", "tier": "frontier", "status": "verified", "input": "15.00", "output": "75.00", "blended": "30.0000", "source": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "page": "https://docs.anthropic.com/claude/docs/models-overview", "fetchedAt": "2026-10-19T22:02:10+00:00", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "snapshotAt": "2024-04-17T14:49:03+00:00", "sha256": "06e6a29c2b9059eedc58f4ce091ba4b6250ce34dcb4b912453853196b1030bd5", "excerpt": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks Ideal balance of intelligence and speed for enterprise workloads Fastest and most compact model for near-instant responsiveness Updated version of Claude 2 with improved accuracy Predecessor to Claude 3, offering strong all-round performance Our cheapest small and fast model, a predecessor of Claude Haiku. Strengths Top-level performance, intelligence, fluency, and understanding Maximum utility at a lower price, dependable, balanced for scaled deployments Quick and accurate targeted performance Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Multilingual Yes Yes Yes Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Vision Yes Yes Yes No No No Latest API model name claude-3-opus-20240229 claude-3-sonnet-20240229 claude-3-haiku-20240307 claude-2.1 claude-2.0 claude-instant-1.2 API format Messages API Messages API Messages API Messages & Text Completions API Messages & Text Completions API Messages & Text Completions API Comparative latency Moderately fast Fast Fastest Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Context window 200K* 200K* 200K* 200K* 100K** 100K** Max output 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens Cost (Input / Output per MTok^) $15.00 / $75.00" }, "positioning": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "quote": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks", "rule": null } }, { "id": "anthropic-claude-3-sonnet", "provider": "Anthropic", "model": "Claude 3 Sonnet", "modelId": "claude-3-sonnet-20240229", "tier": "mid", "status": "verified", "input": "3.00", "output": "15.00", "blended": "6.0000", "source": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "page": "https://docs.anthropic.com/claude/docs/models-overview", "fetchedAt": "2026-10-19T22:02:10+00:00", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "snapshotAt": "2024-04-17T14:49:03+00:00", "sha256": "06e6a29c2b9059eedc58f4ce091ba4b6250ce34dcb4b912453853196b1030bd5", "excerpt": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks Ideal balance of intelligence and speed for enterprise workloads Fastest and most compact model for near-instant responsiveness Updated version of Claude 2 with improved accuracy Predecessor to Claude 3, offering strong all-round performance Our cheapest small and fast model, a predecessor of Claude Haiku. Strengths Top-level performance, intelligence, fluency, and understanding Maximum utility at a lower price, dependable, balanced for scaled deployments Quick and accurate targeted performance Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Multilingual Yes Yes Yes Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Vision Yes Yes Yes No No No Latest API model name claude-3-opus-20240229 claude-3-sonnet-20240229 claude-3-haiku-20240307 claude-2.1 claude-2.0 claude-instant-1.2 API format Messages API Messages API Messages API Messages & Text Completions API Messages & Text Completions API Messages & Text Completions API Comparative latency Moderately fast Fast Fastest Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Context window 200K* 200K* 200K* 200K* 100K** 100K** Max output 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens Cost (Input / Output per MTok^) $15.00 / $75.00 $3.00 / $15.00" }, "positioning": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "quote": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks Ideal balance of intelligence and speed for enterprise workloads", "rule": null } }, { "id": "anthropic-claude-3-haiku", "provider": "Anthropic", "model": "Claude 3 Haiku", "modelId": "claude-3-haiku-20240307", "tier": "small", "status": "verified", "input": "0.25", "output": "1.25", "blended": "0.5000", "source": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "page": "https://docs.anthropic.com/claude/docs/models-overview", "fetchedAt": "2026-10-19T22:02:10+00:00", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "snapshotAt": "2024-04-17T14:49:03+00:00", "sha256": "06e6a29c2b9059eedc58f4ce091ba4b6250ce34dcb4b912453853196b1030bd5", "excerpt": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks Ideal balance of intelligence and speed for enterprise workloads Fastest and most compact model for near-instant responsiveness Updated version of Claude 2 with improved accuracy Predecessor to Claude 3, offering strong all-round performance Our cheapest small and fast model, a predecessor of Claude Haiku. Strengths Top-level performance, intelligence, fluency, and understanding Maximum utility at a lower price, dependable, balanced for scaled deployments Quick and accurate targeted performance Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Legacy model - performs less well than Claude 3 models Multilingual Yes Yes Yes Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Yes, with less coverage, understanding, and skill than Claude 3 Vision Yes Yes Yes No No No Latest API model name claude-3-opus-20240229 claude-3-sonnet-20240229 claude-3-haiku-20240307 claude-2.1 claude-2.0 claude-instant-1.2 API format Messages API Messages API Messages API Messages & Text Completions API Messages & Text Completions API Messages & Text Completions API Comparative latency Moderately fast Fast Fastest Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Slower than Claude 3 model of similar intelligence Context window 200K* 200K* 200K* 200K* 100K** 100K** Max output 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens 4096 tokens Cost (Input / Output per MTok^) $15.00 / $75.00 $3.00 / $15.00 $0.25 / $1.25" }, "positioning": { "url": "https://docs.anthropic.com/claude/docs/models-overview", "snapshotUrl": "https://web.archive.org/web/20240417144903/https://docs.anthropic.com/claude/docs/models-overview", "quote": "Claude 3 Opus Claude 3 Sonnet Claude 3 Haiku Claude 2.1 Claude 2 Claude Instant 1.2 Description Most powerful model for highly complex tasks Ideal balance of intelligence and speed for enterprise workloads Fastest and most compact model for near-instant responsiveness", "rule": null } }, { "id": "mistral-large", "provider": "Mistral AI", "model": "Mistral Large", "modelId": "mistral-large-2402", "tier": "frontier", "status": "verified", "input": "8", "output": "24", "blended": "12.0000", "source": { "url": "https://mistral.ai/technology", "page": "https://mistral.ai/technology", "fetchedAt": "2026-10-19T22:09:18+00:00", "snapshotUrl": "https://web.archive.org/web/20240425211109/https://mistral.ai/technology/", "snapshotAt": "2024-04-25T21:11:09+00:00", "sha256": "e3af51b862a9d94d47a075bf9401b6a9fb75ce4c73f9d5c9f695ca0de39e0ccf", "excerpt": "mistral-large Top-tier reasoning for high-complexity tasks. The most powerful model of the Mistral AI family. $8/1M tokens $24" }, "positioning": { "url": "https://mistral.ai/technology", "snapshotUrl": "https://web.archive.org/web/20240425211109/https://mistral.ai/technology/", "quote": "mistral-large Top-tier reasoning for high-complexity tasks. The most powerful model of the Mistral AI family.", "rule": "Mistral Medium was sold on this date but Mistral did not present it as a model balancing capability with speed or cost (it called it \"Will soon be deprecated\" or \"Our first commercial model\"), so Mistral has no mid-range constituent." } }, { "id": "mistral-small", "provider": "Mistral AI", "model": "Mistral Small", "modelId": "mistral-small-2402", "tier": "small", "status": "verified", "input": "2", "output": "6", "blended": "3.0000", "source": { "url": "https://mistral.ai/technology", "page": "https://mistral.ai/technology", "fetchedAt": "2026-10-19T22:09:18+00:00", "snapshotUrl": "https://web.archive.org/web/20240425211109/https://mistral.ai/technology/", "snapshotAt": "2024-04-25T21:11:09+00:00", "sha256": "e3af51b862a9d94d47a075bf9401b6a9fb75ce4c73f9d5c9f695ca0de39e0ccf", "excerpt": "mistral-small Cost-efficient reasoning for low-latency workloads. $2/1M tokens $6" }, "positioning": { "url": "https://mistral.ai/technology", "snapshotUrl": "https://web.archive.org/web/20240425211109/https://mistral.ai/technology/", "quote": "mistral-small Cost-efficient reasoning for low-latency workloads.", "rule": "Mistral's open-weight models (Mistral 7B, Mixtral, NeMo) are cheaper, but the model Mistral itself positions for low-latency or high-volume work is Mistral Small, which takes the small slot." } } ] }, "editions": [ { "edition": "2024-01", "label": "January 2024", "url": "https://kingy.ai/ai-price-index/data/2024-01.json" }, { "edition": "2024-02", "label": "February 2024", "url": "https://kingy.ai/ai-price-index/data/2024-02.json" }, { "edition": "2024-03", "label": "March 2024", "url": "https://kingy.ai/ai-price-index/data/2024-03.json" }, { "edition": "2024-04", "label": "April 2024", "url": "https://kingy.ai/ai-price-index/data/2024-04.json" }, { "edition": "2024-05", "label": "May 2024", "url": "https://kingy.ai/ai-price-index/data/2024-05.json" }, { "edition": "2024-06", "label": "June 2024", "url": "https://kingy.ai/ai-price-index/data/2024-06.json" }, { "edition": "2024-07", "label": "July 2024", "url": "https://kingy.ai/ai-price-index/data/2024-07.json" }, { "edition": "2024-08", "label": "August 2024", "url": "https://kingy.ai/ai-price-index/data/2024-08.json" }, { "edition": "2024-09", "label": "September 2024", "url": "https://kingy.ai/ai-price-index/data/2024-09.json" }, { "edition": "2024-10", "label": "October 2024", "url": "https://kingy.ai/ai-price-index/data/2024-10.json" }, { "edition": "2024-11", "label": "November 2024", "url": "https://kingy.ai/ai-price-index/data/2024-11.json" }, { "edition": "2024-12", "label": "December 2024", "url": "https://kingy.ai/ai-price-index/data/2024-12.json" }, { "edition": "2025-01", "label": "January 2025", "url": "https://kingy.ai/ai-price-index/data/2025-01.json" }, { "edition": "2025-02", "label": "February 2025", "url": "https://kingy.ai/ai-price-index/data/2025-02.json" }, { "edition": "2025-03", "label": "March 2025", "url": "https://kingy.ai/ai-price-index/data/2025-03.json" }, { "edition": "2025-04", "label": "April 2025", "url": "https://kingy.ai/ai-price-index/data/2025-04.json" }, { "edition": "2025-05", "label": "May 2025", "url": "https://kingy.ai/ai-price-index/data/2025-05.json" }, { "edition": "2025-06", "label": "June 2025", "url": "https://kingy.ai/ai-price-index/data/2025-06.json" }, { "edition": "2025-07", "label": "July 2025", "url": "https://kingy.ai/ai-price-index/data/2025-07.json" }, { "edition": "2025-08", "label": "August 2025", "url": "https://kingy.ai/ai-price-index/data/2025-08.json" }, { "edition": "2025-09", "label": "September 2025", "url": "https://kingy.ai/ai-price-index/data/2025-09.json" }, { "edition": "2025-10", "label": "October 2025", "url": "https://kingy.ai/ai-price-index/data/2025-10.json" }, { "edition": "2025-11", "label": "November 2025", "url": "https://kingy.ai/ai-price-index/data/2025-11.json" }, { "edition": "2025-12", "label": "December 2025", "url": "https://kingy.ai/ai-price-index/data/2025-12.json" }, { "edition": "2026-01", "label": "January 2026", "url": "https://kingy.ai/ai-price-index/data/2026-01.json" }, { "edition": "2026-02", "label": "February 2026", "url": "https://kingy.ai/ai-price-index/data/2026-02.json" }, { "edition": "2026-03", "label": "March 2026", "url": "https://kingy.ai/ai-price-index/data/2026-03.json" }, { "edition": "2026-04", "label": "April 2026", "url": "https://kingy.ai/ai-price-index/data/2026-04.json" }, { "edition": "2026-05", "label": "May 2026", "url": "https://kingy.ai/ai-price-index/data/2026-05.json" }, { "edition": "2026-06", "label": "June 2026", "url": "https://kingy.ai/ai-price-index/data/2026-06.json" }, { "edition": "2026-07", "label": "July 2026", "url": "https://kingy.ai/ai-price-index/data/2026-07.json" }, { "edition": "2026-08", "label": "August 2026", "url": "https://kingy.ai/ai-price-index/data/2026-08.json" }, { "edition": "2026-09", "label": "September 2026", "url": "https://kingy.ai/ai-price-index/data/2026-09.json" }, { "edition": "2026-10", "label": "October 2026", "url": "https://kingy.ai/ai-price-index/data/2026-10.json" } ] }