# AI Model API Pricing: Anthropic, OpenAI, Google, xAI and More
Source: https://savrn.com/ai-index/pricing/models
Markdown alternate of the page above; the site index is https://savrn.com/llms.txt

---

[SAVRN Index.](https://savrn.com/ai-index)

Explore the IndexAI pricing

AI pricing pagesFrontier model pricing

Tracking 796 prices · Checked 2 hours ago

Frontier models, listed prices Unchanged

No model has changed a listed price since Sep 11, 2026. 6 new models launched from the labs in that time, each priced against the one it replaces in [price changes](https://savrn.com/ai-index/pricing/changes#launches). The typical output per million tokens is $3.63 ($3.34 on Sep 11, 2026); it moves as models launch and retire, not because anyone repriced.

[Full price history →](https://savrn.com/ai-index/pricing/history#model)

Same model, different seller

## The same model, priced by each seller

Some models are sold by the lab and by a cloud marketplace. Each card is one model: every seller's output price per million tokens at the standard tier, lowest first, and how much more the others charge.

Grok 4.710% apart

Microsoft Azure $6.00Lowest

xAI $6.00Same

AWS Bedrock $6.60+10%

Output per million tokens. Input from $2.00

Claude Fable 510% apart

Anthropic $50.00Lowest

AWS Bedrock $55.00+10%

Output per million tokens. Input from $10.00

Claude Fable 5.110% apart

Anthropic $50.00Lowest

AWS Bedrock $55.00+10%

Output per million tokens. Input from $10.00

Claude Mythos 510% apart

Anthropic $50.00Lowest

AWS Bedrock $55.00+10%

Output per million tokens. Input from $10.00

DeepSeek V3.1194% apart

AWS Bedrock $1.68Lowest

Microsoft Azure $4.94+194%

Output per million tokens. Input from $0.58

DeepSeek V3.210.1% apart

Microsoft Azure $1.68Lowest

AWS Bedrock $1.85+10%

Output per million tokens. Input from $0.58

Llama 4 Maverick18.6% apart

AWS Bedrock $0.97Lowest

Google Vertex $1.15+19%

Output per million tokens. Input from $0.24

Llama 3.3 70B1.4% apart

Microsoft Azure $0.71Lowest

AWS Bedrock $0.72+1%

Output per million tokens. Input from $0.71

GPT-5.4 ProSame price

Microsoft Azure $180.00Same

OpenAI $180.00Same

Output per million tokens. Input from $30.00

GPT-5.2 ProSame price

Microsoft Azure $168.00Same

OpenAI $168.00Same

Output per million tokens. Input from $21.00

## Anthropic

15 models, output $5.00 to $50.00 per million tokens Newest models first · 2 new Checked Oct 24, 2026 · [platform.claude.com ↗](https://platform.claude.com/docs/en/about-claude/pricing)

| Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Claude Fable 5.1Cache hits and refreshes on Claude Fable 5.1 and Claude Mythos 5.1 are priced at 0.025x the base input price. | $10.00 | $0.25 | $50.00 | $111.03 | Batch $5.00 / $25.00 | No change | On pageOct 24, 2026 |
| Claude Mythos 5.1limited availabilityCache hits and refreshes on Claude Fable 5.1 and Claude Mythos 5.1 are priced at 0.025x the base input price. | $10.00 | $0.25 | $50.00 | $111.03 | Batch $5.00 / $25.00 | No change | On pageOct 24, 2026 |
| Claude Fable 5 | $10.00 | $1.00 | $50.00 | $111.03 | Batch $5.00 / $25.00 | No change | On pageOct 24, 2026 |
| Claude Mythos 5limited availability | $10.00 | $1.00 | $50.00 | $111.03 | Batch $5.00 / $25.00 | No change | On pageOct 24, 2026 |
| Claude Opus 5.5Cache hits and refreshes on Claude Opus 5.5 and Claude Sonnet 5.5 are priced at 0.05x the base input price. | $4.00 | $0.20 | $20.00 | $44.41 | Batch $2.00 / $10.00 | New · Oct 5 | On pageOct 24, 2026 |
| Claude Opus 5 | $5.00 | $0.50 | $25.00 | $55.51 | Batch $2.50 / $12.50 | No change | On pageOct 24, 2026 |
| Claude Opus 4.8 | $5.00 | $0.50 | $25.00 | $55.51 | Batch $2.50 / $12.50 | No change | On pageOct 24, 2026 |
| Claude Opus 4.7 | $5.00 | $0.50 | $25.00 | $55.51 | Batch $2.50 / $12.50 | No change | On pageOct 24, 2026 |
| Claude Opus 4.6 | $5.00 | $0.50 | $25.00 | $55.51 | Batch $2.50 / $12.50 | No change | On pageOct 24, 2026 |
| Claude Opus 4.5 | $5.00 | $0.50 | $25.00 | $55.51 | Batch $2.50 / $12.50 | No change | On pageOct 24, 2026 |
| Claude Sonnet 5.5Cache hits and refreshes on Claude Opus 5.5 and Claude Sonnet 5.5 are priced at 0.05x the base input price. | $2.00 | $0.20 | $10.00 | $22.21 | Batch $1.00 / $5.00 | New · Sep 29 | RecheckingOct 24, 2026 |
| Claude Sonnet 5The $2/$10 per million input/output token pricing for Claude Sonnet 5, announced at launch as introductory pricing through August 31, 2026, is now the standard price. | $2.00 | $0.20 | $10.00 | $22.21 | Batch $1.00 / $5.00 | No change | On pageOct 24, 2026 |
| Claude Sonnet 4.6 | $3.00 | $0.30 | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | On pageOct 24, 2026 |
| Claude Sonnet 4.5 | $3.00 | $0.30 | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | On pageOct 24, 2026 |
| Claude Haiku 4.5 | $1.00 | $0.10 | $5.00 | $11.10 | Batch $0.50 / $2.50 | No change | On pageOct 24, 2026 |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

Committed capacity

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| Anthropic | Priority TierSold per Input TPM + output TPM commitment | n.a., not published | 1 / 3 / 6 / 12 monthsCache read 0.1×, 5-min write 1.25×, 1-hr write 2.0× weighting | ResearchOct 21, 2026[docs.anthropic.com ↗](https://docs.anthropic.com/en/api/service-tiers) |

## OpenAI

40 models, output $0.40 to $600.00 per million tokens Newest models first · 3 new Checked Oct 19, 2026 · [developers.openai.com ↗](https://developers.openai.com/api/docs/pricing)

| Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- |
| GPT-6 AstraShort context cache writes $12.50Long context cache writes $25.00 | $10.00 | $1.00 | $50.00 | $111.03 | Batch $5.00 / $25.00Flex $5.00 / $25.00Fast $20.00 / $100.00Ultrafast $60.00 / $300.00Over 272K input tokens: $20.00 / $75.00 | No change | On pageOct 19, 2026 |
| GPT-6.1 SolShort context cache writes $2.50Long context cache writes $5.00 | $2.00 | $0.10 | $10.00 | $22.21 | Batch $1.00 / $5.00Flex $1.00 / $5.00Fast $4.00 / $20.00Ultrafast $12.00 / $60.00Over 272K input tokens: $4.00 / $15.00 | New · Sep 30 | On pageOct 19, 2026 |
| GPT-6 LunaShort context cache writes $0.125Long context cache writes $0.25 | $0.10 | $0.01 | $0.50 | $1.11 | Batch $0.05 / $0.25Flex $0.05 / $0.25Fast $0.20 / $1.00Over 272K input tokens: $0.20 / $0.75 | New · Sep 23 | On pageOct 19, 2026 |
| GPT-6 SolShort context cache writes $2.50Long context cache writes $5.00 | $2.00 | $0.20 | $10.00 | $22.21 | Batch $1.00 / $5.00Flex $1.00 / $5.00Fast $4.00 / $20.00Over 272K input tokens: $4.00 / $15.00 | New · Sep 23 | On pageOct 19, 2026 |
| GPT-5.6 Solpromo to 21 Nov 2026Short context cache writes $5.00Long context cache writes $10.00 | $4.00 | $0.40 | $20.00 | $44.41 | Batch $2.00 / $10.00Flex $2.00 / $10.00Fast $8.00 / $40.00Over 272K input tokens: $8.00 / $30.00 | No change | On pageOct 19, 2026 |
| GPT-5.6 TerraShort context cache writes $2.50Long context cache writes $5.00 | $2.00 | $0.20 | $12.00 | $23.91 | Batch $1.00 / $6.00Flex $1.00 / $6.00Fast $4.00 / $24.00Over 272K input tokens: $4.00 / $18.00 | No change | On pageOct 19, 2026 |
| GPT-5.6 LunaShort context cache writes $0.25Long context cache writes $0.50 | $0.20 | $0.02 | $1.20 | $2.39 | Batch $0.10 / $0.60Flex $0.10 / $0.60Fast $0.40 / $2.40Over 272K input tokens: $0.40 / $1.80 | No change | On pageOct 19, 2026 |
| GPT-5.5 | $5.00 | $0.50 | $30.00 | $59.78 | Batch $2.50 / $15.00Flex $2.50 / $15.00Fast $12.50 / $75.00Over 272K input tokens: $10.00 / $45.00 | No change | On pageOct 19, 2026 |
| GPT-5.5 Pro | $30.00 | n.a. | $180.00 | $358.71 | Batch $15.00 / $90.00Flex $15.00 / $90.00Over 272K input tokens: $60.00 / $270.00 | No change | On pageOct 19, 2026 |
| GPT-5.4 | $2.50 | $0.25 | $15.00 | $29.89 | Batch $1.25 / $7.50Flex $1.25 / $7.50Fast $5.00 / $30.00Over 272K input tokens: $5.00 / $22.50 | No change | On pageOct 19, 2026 |
| GPT-5.4 Mini | $0.75 | $0.075 | $4.50 | $8.97 | Batch $0.38 / $2.25Flex $0.38 / $2.25Fast $1.50 / $9.00 | No change | On pageOct 19, 2026 |
| GPT-5.4 Nano | $0.20 | $0.02 | $1.25 | $2.43 | Batch $0.10 / $0.63Flex $0.10 / $0.63 | No change | On pageOct 19, 2026 |
| GPT-5.4 Pro | $30.00 | n.a. | $180.00 | $358.71 | Batch $15.00 / $90.00Flex $15.00 / $90.00Over 272K input tokens: $60.00 / $270.00 | No change | On pageOct 19, 2026 |
| GPT-5.2 | $1.75 | $0.18 | $14.00 | $23.91 | Batch $0.88 / $7.00Flex $0.88 / $7.00Fast $3.50 / $28.00 | No change | On pageOct 19, 2026 |
| GPT-5.2 Pro | $21.00 | n.a. | $168.00 | $286.97 | Batch $10.50 / $84.00 | No change | On pageOct 19, 2026 |
| GPT-5.1 | $1.25 | $0.13 | $10.00 | $17.08 | Batch $0.63 / $5.00Flex $0.63 / $5.00Fast $2.50 / $20.00 | No change | On pageOct 19, 2026 |
| GPT-5 | $1.25 | $0.13 | $10.00 | $17.08 | Batch $0.63 / $5.00Flex $0.63 / $5.00Fast $2.50 / $20.00 | No change | On pageOct 19, 2026 |
| GPT-5 Mini | $0.25 | $0.025 | $2.00 | $3.42 | Batch $0.13 / $1.00Flex $0.13 / $1.00Fast $0.45 / $3.60 | No change | On pageOct 19, 2026 |
| GPT-5 Nano | $0.05 | $0.005 | $0.40 | $0.68 | Batch $0.025 / $0.20Flex $0.025 / $0.20 | No change | On pageOct 19, 2026 |
| GPT-5 Pro | $15.00 | n.a. | $120.00 | $204.98 | Batch $7.50 / $60.00 | No change | On pageOct 19, 2026 |
| GPT-4.1 | $2.00 | $0.50 | $8.00 | $20.50 | Batch $1.00 / $4.00Fast $3.50 / $14.00 | No change | On pageOct 19, 2026 |
| GPT-4.1 Mini | $0.40 | $0.10 | $1.60 | $4.10 | Batch $0.20 / $0.80Fast $0.70 / $2.80 | No change | On pageOct 19, 2026 |
| GPT-4.1 Nano | $0.10 | $0.025 | $0.40 | $1.02 | Batch $0.05 / $0.20Fast $0.20 / $0.80 | No change | On pageOct 19, 2026 |
| GPT-4o | $2.50 | $1.25 | $10.00 | $25.62 | Batch $1.25 / $5.00Fast $4.25 / $17.00 | No change | On pageOct 19, 2026 |
| GPT-4o 2024-05-13 | $5.00 | n.a. | $15.00 | $46.97 | Batch $2.50 / $7.50Fast $8.75 / $26.25 | No change | On pageOct 19, 2026 |
| GPT-4o Mini | $0.15 | $0.075 | $0.60 | $1.54 | Batch $0.075 / $0.30Fast $0.25 / $1.00 | No change | On pageOct 19, 2026 |
| o1 | $15.00 | $7.50 | $60.00 | $153.73 | Batch $7.50 / $30.00 | No change | On pageOct 19, 2026 |
| o1-pro | $150.00 | n.a. | $600.00 | $1,537 | Batch $75.00 / $300.00 | No change | On pageOct 19, 2026 |
| o3-pro | $20.00 | n.a. | $80.00 | $204.98 | Batch $10.00 / $40.00 | No change | On pageOct 19, 2026 |
| o3 | $2.00 | $0.50 | $8.00 | $20.50 | Batch $1.00 / $4.00Flex $1.00 / $4.00Fast $3.50 / $14.00 | No change | On pageOct 19, 2026 |
| o4-mini | $1.10 | $0.28 | $4.40 | $11.27 | Batch $0.55 / $2.20Flex $0.55 / $2.20Fast $2.00 / $8.00 | No change | On pageOct 19, 2026 |
| o3-mini | $1.10 | $0.55 | $4.40 | $11.27 | Batch $0.55 / $2.20 | No change | On pageOct 19, 2026 |
| GPT-4 Turbo 2024-04-09 | $10.00 | n.a. | $30.00 | $93.95 | Batch $5.00 / $15.00 | No change | On pageOct 19, 2026 |
| GPT-4 0613 | $30.00 | n.a. | $60.00 | $256.22 | Batch $15.00 / $30.00 | No change | On pageOct 19, 2026 |
| GPT-3.5 Turbo | $0.50 | n.a. | $1.50 | $4.70 | None published | No change | On pageOct 19, 2026 |
| GPT-3.5 Turbo 0125 | $0.50 | n.a. | $1.50 | $4.70 | Batch $0.25 / $0.75 | No change | On pageOct 19, 2026 |
| GPT-3.5 Turbo 1106 | $1.00 | n.a. | $2.00 | $8.54 | Batch $1.00 / $2.00 | No change | On pageOct 19, 2026 |
| GPT-3.5 Turbo Instruct | $1.50 | n.a. | $2.00 | $11.96 | None published | No change | On pageOct 19, 2026 |
| davinci-002 | $2.00 | n.a. | $2.00 | $15.37 | Batch $1.00 / $1.00 | No change | On pageOct 19, 2026 |
| babbage-002 | $0.40 | n.a. | $0.40 | $3.07 | Batch $0.20 / $0.20 | No change | On pageOct 19, 2026 |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

Committed capacity

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| OpenAI | Reserved TierSold per Dollars of list-price usage per minute | n.a., enterprise only | n.a.Excludes Batch and Flex | ResearchOct 21, 2026[openai.com ↗](https://openai.com/api-reserved-tier/) |
| OpenAI | Scale Tier (GPT-5.5)Sold per 50,000 input TPM unit per day | $750.00 per 50,000 input TPM unit per day$750 / unit / day | 30-day minimumCommitted tokens-per-minute; GPT-5.4 $300, 5.4-mini $100 per unit/day | ResearchOct 21, 2026[openai.com ↗](https://openai.com/api-scale-tier/) |

## Google

11 models, output $0.40 to $12.00 per million tokens Newest models first Checked Oct 24, 2026 · [ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) · [cloud.google.com ↗](https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing)

| Seller | Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- | --- |
| Google | Gemini 3.8 FlashCurrent prices apply through December 31, 2026Context caching storage $0.50 per 1M tokens per hour through December 31, 2026 | $0.75 | $0.075 | $3.75 | $8.33 | Batch $0.38 / $1.88Flex $0.38 / $1.88Priority $1.35 / $6.75From Jan 1, 2027: $1.50 / $7.50 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 3.6 FlashCurrent prices apply through December 31, 2026Context caching storage $0.50 per 1M tokens per hour through December 31, 2026 | $0.75 | $0.075 | $3.75 | $8.33 | Batch $0.38 / $1.88Flex $0.38 / $1.88Priority $1.35 / $6.75From Jan 1, 2027: $1.50 / $7.50 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 3.5 Flash-LiteContext caching storage $1.00 per 1M tokens per hour | $0.30 | $0.03 | $2.50 | $4.18 | Batch $0.15 / $1.25Flex $0.15 / $1.25Priority $0.54 / $4.50 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 3.1 Flash-LiteContext caching storage $1.00 per 1M tokens per hour | $0.25 | $0.025 | $1.50 | $2.99 | Batch $0.13 / $0.75Flex $0.13 / $0.75Priority $0.45 / $2.70 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 3.1 Pro Preview≤200KContext caching storage $4.50 per 1M tokens per hour | $2.00 | $0.20 | $12.00 | $23.91 | Batch $1.00 / $6.00Flex $1.00 / $6.00Priority $3.60 / $21.60Over 200K input tokens: $4.00 / $18.00 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 3 Flash PreviewContext caching storage $1.00 per 1M tokens per hour | $0.50 | $0.05 | $3.00 | $5.98 | Batch $0.25 / $1.50Flex $0.25 / $1.50Priority $0.90 / $5.40 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 2.5 ProContext caching storage $4.50 per 1M tokens per hour | $1.25 | $0.13 | $10.00 | $17.08 | Batch $0.63 / $5.00Flex $0.63 / $5.00Priority $2.25 / $18.00Over 200K input tokens: $2.50 / $15.00 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 2.5 FlashContext caching storage $1.00 per 1M tokens per hour | $0.30 | $0.03 | $2.50 | $4.18 | Batch $0.15 / $1.25Flex $0.15 / $1.25Priority $0.54 / $4.50 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini 2.5 Flash-LiteContext caching storage $1.00 per 1M tokens per hour | $0.10 | $0.01 | $0.40 | $1.02 | Batch $0.05 / $0.20Flex $0.05 / $0.20Priority $0.18 / $0.72 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google | Gemini Robotics ER 2 PreviewCurrent prices apply through December 31, 2026Context caching storage $0.50 per 1M tokens per hour through December 31, 2026 | $1.00 | $0.10 | $5.00 | $11.10 | Batch $0.50 / $2.50From Jan 1, 2027: $2.00 / $10.00 | No change | On pageOct 24, 2026[ai.google.dev ↗](https://ai.google.dev/gemini-api/docs/pricing) |
| Google Vertex | Llama 4 Maverick | $0.35 | n.a. | $1.15 | $3.37 | None published | No change | ResearchOct 21, 2026[cloud.google.com ↗](https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing) |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

Committed capacity

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| Google Vertex AI | Provisioned Throughput (Global)Sold per GSU-hour | $7.14 (1-wk) → $3.70 (1-mo) → $3.29 (3-mo) → $2.74 (1-yr) | 1 week to 1 year; 1-yr vs 1-mo −25.9%Gemini 2.5 Flash: 2,690 tok/s per GSU; Gemini 2.0 Flash: 3,360 tok/s per GSU | ResearchOct 21, 2026[cloud.google.com ↗](https://cloud.google.com/gemini-enterprise-agent-platform/generative-ai/pricing) |

## xAI

8 models, output $2.00 to $6.00 per million tokens Newest models first · 1 new Checked Oct 24, 2026 · [docs.x.ai ↗](https://docs.x.ai/developers/models)

| Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Grok 4.7Batch discount 20%Priced up to 10% higher in some regions (us-central-1) | $2.00 | $0.50 | $6.00 | $18.79 | Over 200K input tokens: $4.00 / $12.00 | New · Oct 5 | On pageOct 24, 2026 |
| Grok 4.6Batch discount 20%Priced up to 10% higher in some regions (us-central-1) | $2.00 | $0.50 | $6.00 | $18.79 | Over 200K input tokens: $4.00 / $12.00 | No change | On pageOct 24, 2026 |
| Grok 4.5Batch discount 20% | $2.00 | $0.30 | $6.00 | $18.79 | Over 200K input tokens: $4.00 / $12.00 | No change | On pageOct 24, 2026 |
| Grok 4.3Batch discount 20% | $1.25 | $0.20 | $2.50 | $10.68 | Over 200K input tokens: $2.50 / $5.00 | No change | On pageOct 24, 2026 |
| Grok 4.20 0309 ReasoningBatch discount 20% | $1.25 | $0.20 | $2.50 | $10.68 | Over 200K input tokens: $2.50 / $5.00 | No change | On pageOct 24, 2026 |
| Grok 4.20 Multi-Agent 0309Batch discount 20% | $1.25 | $0.20 | $2.50 | $10.68 | Over 200K input tokens: $2.50 / $5.00 | No change | On pageOct 24, 2026 |
| Grok 4.20 0309 Non-ReasoningBatch discount 20% | $1.25 | $0.20 | $2.50 | $10.68 | Over 200K input tokens: $2.50 / $5.00 | No change | On pageOct 24, 2026 |
| Grok Build 0.1Batch discount 20% | $1.00 | $0.20 | $2.00 | $8.54 | Over 200K input tokens: $2.00 / $4.00 | No change | On pageOct 24, 2026 |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

## Microsoft Azure

83 models, output $0.30 to $180.00 per million tokens Checked Oct 24, 2026 · [prices.azure.com ↗](https://prices.azure.com/api/retail/prices)

| Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- |
| GPT-5.4 ProModel by OpenAI | $30.00 | n.a. | $180.00 | $358.71 | Batch $15.00 / $90.00Flex $15.00 / $90.00Data Zone $33.00 / $198.00Over the long-context threshold: $60.00 / $270.00 | No change | FeedOct 24, 2026 |
| GPT-5.2 ProModel by OpenAI | $21.00 | n.a. | $168.00 | $286.97 | Batch $10.50 / $84.00Data Zone $23.10 / $184.80 | No change | FeedOct 24, 2026 |
| GPT-5 ProModel by OpenAI | $15.00 | n.a. | $120.00 | $204.98 | Batch $7.50 / $60.00Data Zone $16.50 / $132.00 | No change | FeedOct 24, 2026 |
| GPT-6 AstraModel by OpenAI | $10.00 | $1.00 | $50.00 | $111.03 | Flex $5.00 / $25.00Data Zone $11.00 / $55.00Over the long-context threshold: $20.00 / $75.00 | No change | FeedOct 24, 2026 |
| GPT-5.5Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | Batch $2.50 / $15.00Priority $12.50 / $75.00Data Zone $5.50 / $33.00Over the long-context threshold: $10.00 / $45.00 | No change | FeedOct 24, 2026 |
| chat-latest 10072026Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | Data Zone $5.50 / $33.00 | New · Oct 22 | FeedOct 24, 2026 |
| chat-latest 08062026Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | Data Zone $5.50 / $33.00 | No change | FeedOct 24, 2026 |
| chat-latest 07012026Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | Data Zone $5.50 / $33.00 | No change | FeedOct 24, 2026 |
| chat-latest 05282026Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | None published | No change | FeedOct 24, 2026 |
| chat-latest 05052026Model by OpenAI | $5.00 | $0.50 | $30.00 | $59.78 | None published | No change | FeedOct 24, 2026 |
| GPT-5.6 SolModel by OpenAI | $4.00 | $0.40 | $20.00 | $44.41 | Flex $2.00 / $10.00Priority $8.00 / $40.00Data Zone $4.40 / $22.00Over the long-context threshold: $8.00 / $30.00 | No change | FeedOct 24, 2026 |
| Grok 4Model by xAI | $3.00 | n.a. | $15.00 | $33.31 | Data Zone $3.30 / $16.50 | No change | FeedOct 24, 2026 |
| Grok 3Model by xAI | $3.00 | n.a. | $15.00 | $33.31 | Data Zone $3.30 / $16.50 | No change | FeedOct 24, 2026 |
| GPT-5.4Model by OpenAI | $2.50 | $0.25 | $15.00 | $29.89 | Batch $1.25 / $7.50Flex $1.25 / $7.50Priority $5.00 / $30.00Data Zone $2.75 / $16.50Over the long-context threshold: $5.00 / $22.50 | No change | FeedOct 24, 2026 |
| GPT-5.3 CodexModel by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Priority $3.50 / $28.00Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.3 ChatModel by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.2 CodexModel by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.2 Chat 0210Model by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.2 ChatModel by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.2Model by OpenAI | $1.75 | $0.18 | $14.00 | $23.91 | Batch $0.88 / $7.00Priority $3.50 / $28.00Data Zone $1.93 / $15.40 | No change | FeedOct 24, 2026 |
| GPT-5.6 TerraModel by OpenAI | $2.00 | $0.20 | $12.00 | $23.91 | Flex $1.00 / $6.00Priority $4.00 / $24.00Data Zone $2.20 / $13.20Over the long-context threshold: $4.00 / $18.00 | No change | FeedOct 24, 2026 |
| GPT-6.1 SolModel by OpenAI | $2.00 | $0.10 | $10.00 | $22.21 | Flex $1.00 / $5.00Data Zone $2.20 / $11.00Over the long-context threshold: $4.00 / $15.00 | New · Oct 9 | FeedOct 24, 2026 |
| GPT-6 SolModel by OpenAI | $2.00 | $0.20 | $10.00 | $22.21 | Flex $1.00 / $5.00Priority $4.00 / $20.00Data Zone $2.20 / $11.00Over the long-context threshold: $4.00 / $15.00 | New · Sep 25 | FeedOct 24, 2026 |
| GPT-5.1 Codex MaxModel by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| GPT-5.1 CodexModel by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| GPT-5.1 ChatModel by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| GPT-5.1Model by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Batch $0.63 / $5.00Priority $2.50 / $20.00Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| GPT-5 CodexModel by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| GPT-5 ChatModel by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | None published | No change | FeedOct 24, 2026 |
| GPT-5Model by OpenAI | $1.25 | $0.13 | $10.00 | $17.08 | Batch $0.63 / $5.00Priority $2.50 / $20.00Data Zone $1.38 / $11.00 | No change | FeedOct 24, 2026 |
| MAI-Thinking-1Model by Microsoft | $2.00 | $0.20 | $8.00 | $20.50 | None published | No change | FeedOct 24, 2026 |
| Grok 4.7Model by xAI | $2.00 | $0.50 | $6.00 | $18.79 | Data Zone $2.20 / $6.60Over the long-context threshold: $4.00 / $12.00 | New · Oct 10 | FeedOct 24, 2026 |
| Grok 4.6Model by xAI | $2.00 | $0.50 | $6.00 | $18.79 | Data Zone $2.20 / $6.60Over the long-context threshold: $4.00 / $12.00 | No change | FeedOct 24, 2026 |
| MAI-DS-R1Model by Microsoft | $1.35 | n.a. | $5.40 | $13.84 | Data Zone $1.49 / $5.94 | No change | FeedOct 24, 2026 |
| DeepSeek R1Model by DeepSeek | $1.35 | n.a. | $5.40 | $13.84 | Data Zone $1.49 / $5.94 | No change | FeedOct 24, 2026 |
| DeepSeek V3.1Model by DeepSeek | $1.23 | n.a. | $4.94 | $12.62 | Data Zone $1.36 / $5.43 | No change | FeedOct 24, 2026 |
| DeepSeek V3-0324Model by DeepSeek | $1.14 | n.a. | $4.56 | $11.68 | Data Zone $1.25 / $5.00 | No change | FeedOct 24, 2026 |
| DeepSeek V3Model by DeepSeek | $1.14 | n.a. | $4.56 | $11.68 | Data Zone $1.25 / $5.00 | No change | FeedOct 24, 2026 |
| GPT-5.4 MiniModel by OpenAI | $0.75 | $0.075 | $4.50 | $8.97 | Batch $0.38 / $2.25Flex $0.38 / $2.25Priority $1.50 / $9.00Data Zone $0.83 / $4.95 | No change | FeedOct 24, 2026 |
| Kimi K2.7 CodeModel by Moonshot AI | $0.95 | $0.19 | $4.00 | $9.91 | Data Zone $1.05 / $5.00 | No change | FeedOct 24, 2026 |
| Kimi K2.6 ThinkingModel by Moonshot AI | $0.95 | n.a. | $4.00 | $9.91 | Data Zone $1.05 / $5.00 | No change | FeedOct 24, 2026 |
| MAI-Cyber-1-FlashModel by Microsoft | $0.60 | $0.06 | $3.50 | $7.09 | None published | No change | FeedOct 24, 2026 |
| Cyber 1.1Model by Microsoft | $0.60 | $0.06 | $3.50 | $7.09 | None published | New · Oct 18 | FeedOct 24, 2026 |
| DeepSeek V4 ProModel by DeepSeek | $1.74 | $0.15 | $3.48 | $14.86 | Data Zone $1.91 / $3.83 | No change | FeedOct 24, 2026 |
| Kimi K2.5 ThinkingModel by Moonshot AI | $0.60 | n.a. | $3.00 | $6.66 | Data Zone $0.66 / $3.30 | No change | FeedOct 24, 2026 |
| Kimi K2 ThinkingModel by Moonshot AI | $0.60 | n.a. | $2.50 | $6.23 | Data Zone $0.66 / $2.75 | No change | FeedOct 24, 2026 |
| Grok 4.3Model by xAI | $1.25 | $0.20 | $2.50 | $10.68 | Data Zone $1.38 / $2.75Over the long-context threshold: $2.50 / $5.00 | No change | FeedOct 24, 2026 |
| Grok 4.2Model by xAI | $1.25 | n.a. | $2.50 | $10.68 | Data Zone $1.38 / $2.75 | No change | FeedOct 24, 2026 |
| GPT-5.1 Codex MiniModel by OpenAI | $0.25 | $0.025 | $2.00 | $3.42 | Data Zone $0.28 / $2.20 | No change | FeedOct 24, 2026 |
| GPT-5 MiniModel by OpenAI | $0.25 | $0.025 | $2.00 | $3.42 | Batch $0.13 / $1.00Priority $0.45 / $3.60Data Zone $0.28 / $2.20 | No change | FeedOct 24, 2026 |
| DeepSeek V3.2Model by DeepSeek | $0.58 | n.a. | $1.68 | $5.40 | Data Zone $0.64 / $1.85 | No change | FeedOct 24, 2026 |
| Mistral Large 3Model by Mistral AI | $0.50 | n.a. | $1.50 | $4.70 | Data Zone $0.55 / $1.65 | No change | FeedOct 24, 2026 |
| Grok Code Fast 1Model by xAI | $0.20 | n.a. | $1.50 | $2.65 | Data Zone $0.22 / $1.65 | No change | FeedOct 24, 2026 |
| DeepSeek V4 Flash 0731Model by DeepSeek | $0.44 | $0.014 | $1.32 | $4.13 | Data Zone $0.48 / $1.45 | No change | FeedOct 24, 2026 |
| Grok 3 MiniModel by xAI | $0.25 | n.a. | $1.27 | $2.79 | Data Zone $0.28 / $1.38 | No change | FeedOct 24, 2026 |
| GPT-5.4 NanoModel by OpenAI | $0.20 | $0.02 | $1.25 | $2.43 | Batch $0.10 / $0.63Flex $0.10 / $0.63Data Zone $0.22 / $1.38 | No change | FeedOct 24, 2026 |
| GPT-5.6 LunaModel by OpenAI | $0.20 | $0.02 | $1.20 | $2.39 | Flex $0.10 / $0.60Priority $0.40 / $2.40Data Zone $0.22 / $1.32Over the long-context threshold: $0.40 / $1.80 | No change | FeedOct 24, 2026 |
| DeepSeek V4.1 FlashModel by DeepSeek | $0.30 | $0.006 | $1.20 | $3.07 | Data Zone $0.33 / $1.32 | New · Oct 9 | FeedOct 24, 2026 |
| Code 1.1 FlashModel by Microsoft | $0.20 | $0.02 | $1.20 | $2.39 | None published | New · Sep 17 | FeedOct 24, 2026 |
| Llama 4 Maverick 17BModel by Meta | $0.25 | n.a. | $1.00 | $2.56 | Data Zone $0.28 / $1.10 | No change | FeedOct 24, 2026 |
| CodestralModel by Mistral AI | $0.30 | n.a. | $0.90 | $2.82 | Data Zone $0.33 / $0.99 | No change | FeedOct 24, 2026 |
| Llama 3.3 70BModel by Meta | $0.71 | n.a. | $0.71 | $5.46 | Data Zone $0.78 / $0.78 | No change | FeedOct 24, 2026 |
| Phi-3-Medium-128K-InstructModel by Microsoft | $0.17 | n.a. | $0.68 | $1.74 | None published | No change | FeedOct 24, 2026 |
| Phi-3-Medium-4K-InstructModel by Microsoft | $0.17 | n.a. | $0.68 | $1.74 | None published | No change | FeedOct 24, 2026 |
| Phi-3.5-MoE-InstructModel by Microsoft | $0.16 | n.a. | $0.64 | $1.64 | None published | No change | FeedOct 24, 2026 |
| Phi-3-Small-128K-InstructModel by Microsoft | $0.15 | n.a. | $0.60 | $1.54 | None published | No change | FeedOct 24, 2026 |
| Phi-3-Small-8K-InstructModel by Microsoft | $0.15 | n.a. | $0.60 | $1.54 | None published | No change | FeedOct 24, 2026 |
| gpt-oss-120BModel by OpenAI | $0.15 | n.a. | $0.60 | $1.54 | Data Zone $0.17 / $0.66 | No change | FeedOct 24, 2026 |
| Phi-3.5-V-128K-InstructModel by Microsoft | $0.13 | n.a. | $0.52 | $1.33 | None published | No change | FeedOct 24, 2026 |
| Phi-3.5-Mini-128K-InstructModel by Microsoft | $0.13 | n.a. | $0.52 | $1.33 | None published | No change | FeedOct 24, 2026 |
| Phi-3-Mini-128K-InstructModel by Microsoft | $0.13 | n.a. | $0.52 | $1.33 | None published | No change | FeedOct 24, 2026 |
| Phi-3-Mini-4K-InstructModel by Microsoft | $0.13 | n.a. | $0.52 | $1.33 | None published | No change | FeedOct 24, 2026 |
| DeepSeek V4 FlashModel by DeepSeek | $0.19 | $0.028 | $0.51 | $1.73 | Data Zone $0.21 / $0.56 | No change | FeedOct 24, 2026 |
| Phi-4-reasoning-plusModel by Microsoft | $0.13 | n.a. | $0.50 | $1.28 | None published | No change | FeedOct 24, 2026 |
| Phi-4-reasoningModel by Microsoft | $0.13 | n.a. | $0.50 | $1.28 | None published | No change | FeedOct 24, 2026 |
| Phi-4Model by Microsoft | $0.13 | n.a. | $0.50 | $1.28 | None published | No change | FeedOct 24, 2026 |
| Grok 4.1Model by xAI | $0.20 | n.a. | $0.50 | $1.79 | Data Zone $0.22 / $0.55 | No change | FeedOct 24, 2026 |
| Grok 4 FastModel by xAI | $0.20 | n.a. | $0.50 | $1.79 | Data Zone $0.22 / $0.55 | No change | FeedOct 24, 2026 |
| GPT-6 LunaModel by OpenAI | $0.10 | $0.01 | $0.50 | $1.11 | Flex $0.05 / $0.25Data Zone $0.11 / $0.55Over the long-context threshold: $0.20 / $0.75 | New · Sep 25 | FeedOct 24, 2026 |
| GPT-5 NanoModel by OpenAI | $0.05 | $0.005 | $0.40 | $0.68 | Batch $0.025 / $0.20Data Zone $0.055 / $0.44 | No change | FeedOct 24, 2026 |
| Phi-4-Mini MMModel by Microsoft | $0.08 | n.a. | $0.32 | $0.82 | None published | No change | FeedOct 24, 2026 |
| Phi-4-mini-reasoningModel by Microsoft | $0.075 | n.a. | $0.30 | $0.77 | None published | No change | FeedOct 24, 2026 |
| Phi-4-MiniModel by Microsoft | $0.075 | n.a. | $0.30 | $0.77 | None published | No change | FeedOct 24, 2026 |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

Committed capacity

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| Azure AI Foundry | Provisioned Throughput (Data Zone)Sold per PTU-hour | $1.10 per PTU-hour$1.10 per PTU-hour, Data Zone deployment, eastus2; reservation $286.00 per PTU per month, $2,916.00 per PTU per yearReserved: $286.00 a month, $2,916 a year | Hourly; reservation $286/PTU-month, $2,916/PTU-year | FeedOct 24, 2026[prices.azure.com ↗](https://prices.azure.com/api/retail/prices) |
| Azure AI Foundry | Provisioned Throughput (Global)Sold per PTU-hour | $1.00 per PTU-hour$1.00 per PTU-hour, Global deployment, eastus2; reservation $260.00 per PTU per month, $2,652.00 per PTU per yearReserved: $260.00 a month, $2,652 a year | Hourly; reservation $260/PTU-month, $2,652/PTU-yearModel-dependent TPM per PTU; not published numerically | FeedOct 24, 2026[prices.azure.com ↗](https://prices.azure.com/api/retail/prices) |
| Azure AI Foundry | Provisioned Throughput (Regional, eastus)Sold per PTU-hour | $2.00 per PTU-hour$2.00 per PTU-hour, Regional deployment, eastus; reservation $286.00 per PTU per month, $2,916.00 per PTU per yearReserved: $286.00 a month, $2,916 a year | Hourly; reservation $286/PTU-month, $2,916/PTU-yearMinimum 15 PTUs Global / 50 Regional for GPT-5.x | FeedOct 24, 2026[azure.microsoft.com ↗](https://azure.microsoft.com/en-us/pricing/details/azure-openai/) |

## AWS Bedrock

111 models, output $0.08 to $137.50 per million tokens Checked Oct 24, 2026 · [pricing.us-east-1.amazonaws.com ↗](https://pricing.us-east-1.amazonaws.com/offers/v1.0/aws/AmazonBedrockFoundationModels/current/us-east-1/index.json)

| Model | Input | Cached input | Output | [Revenue per GPU-hour at list](https://savrn.com/ai-index/pricing/gpus/vera-rubin#token-value) | Other published prices | Change | Checked |
| --- | --- | --- | --- | --- | --- | --- | --- |
| Claude Mythos PreviewModel by Anthropic | $27.50 | $2.75 | $137.50 | $305.33 | None published | No change | FeedOct 24, 2026 |
| Claude Opus 4.1Model by Anthropic | $15.00 | $1.50 | $75.00 | $166.54 | None published | No change | FeedOct 24, 2026 |
| Claude Opus 4Model by Anthropic | $15.00 | $1.50 | $75.00 | $166.54 | Batch $7.50 / $37.50 | No change | FeedOct 24, 2026 |
| Claude 3 OpusModel by Anthropic | $15.00 | n.a. | $75.00 | $166.54 | Batch $7.50 / $37.50 | No change | FeedOct 24, 2026 |
| Claude (100K)Model by Anthropic | $22.04 | n.a. | $65.36 | $206.41 | None published | No change | FeedOct 24, 2026 |
| OpenAI GPT-6 Astra | $11.00 | $1.10 | $55.00 | $122.13 | Global $10.00 / $50.00 | New · Sep 28 | FeedOct 24, 2026 |
| Claude Mythos 5.1Model by Anthropic | $11.00 | $0.28 | $55.00 | $122.13 | Global $10.00 / $50.00 | No change | FeedOct 24, 2026 |
| Claude Mythos 5Model by Anthropic | $11.00 | $1.10 | $55.00 | $122.13 | None published | No change | FeedOct 24, 2026 |
| Claude Fable 5.1Model by Anthropic | $11.00 | $0.28 | $55.00 | $122.13 | Global $10.00 / $50.00 | No change | FeedOct 24, 2026 |
| Claude Fable 5Model by Anthropic | $11.00 | $1.10 | $55.00 | $122.13 | Global $10.00 / $50.00 | No change | FeedOct 24, 2026 |
| Claude Opus 5Model by Anthropic | $5.50 | $0.55 | $27.50 | $61.07 | Batch $2.75 / $13.75Global $5.00 / $25.00 | No change | FeedOct 24, 2026 |
| Claude Opus 4.8Model by Anthropic | $5.50 | $0.55 | $27.50 | $61.07 | Global $5.00 / $25.00 | No change | FeedOct 24, 2026 |
| Claude Opus 4.7Model by Anthropic | $5.50 | $0.55 | $27.50 | $61.07 | Global $5.00 / $25.00 | No change | FeedOct 24, 2026 |
| Claude Opus 4.6Model by Anthropic | $5.50 | $0.55 | $27.50 | $61.07 | Batch $2.75 / $13.75Global $5.00 / $25.00 | No change | FeedOct 24, 2026 |
| Claude Opus 4.5Model by Anthropic | $5.50 | $0.55 | $27.50 | $61.07 | Batch $2.75 / $13.75Global $5.00 / $25.00 | No change | FeedOct 24, 2026 |
| ClaudeModel by Anthropic | $8.00 | n.a. | $24.00 | $75.16 | None published | No change | FeedOct 24, 2026 |
| Claude Opus 5.5Model by Anthropic | $4.40 | $0.22 | $22.00 | $48.85 | Global $4.00 / $20.00 | New · Sep 22 | FeedOct 24, 2026 |
| Jurassic-2 UltraModel by AI21 Labs | $18.80 | n.a. | $18.80 | $144.51 | None published | No change | FeedOct 24, 2026 |
| Kimi K3Model by Moonshot AI | $3.30 | $0.33 | $16.50 | $36.64 | Batch $1.65 / $8.25Flex $1.65 / $8.25Priority $5.78 / $28.88Global $3.00 / $15.00 | New · Sep 18 | FeedOct 24, 2026 |
| Claude Sonnet 4.6Model by Anthropic | $3.30 | $0.33 | $16.50 | $36.64 | Batch $1.65 / $8.25Global $3.00 / $15.00 | No change | FeedOct 24, 2026 |
| Claude Sonnet 4.5Model by Anthropic | $3.30 | $0.33 | $16.50 | $36.64 | Batch $1.65 / $8.25Global $3.00 / $15.00 | No change | FeedOct 24, 2026 |
| Cohere Command R+Model by Cohere | $3.00 | n.a. | $15.00 | $33.31 | None published | No change | FeedOct 24, 2026 |
| Claude Sonnet 4Model by Anthropic | $3.00 | $0.30 | $15.00 | $33.31 | Batch $1.50 / $7.50Global $3.00 / $15.00 | No change | FeedOct 24, 2026 |
| Claude 3.7 SonnetModel by Anthropic | $3.00 | $0.30 | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | FeedOct 24, 2026 |
| Claude 3.5 Sonnet v2Model by Anthropic | $3.00 | $0.30 | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | FeedOct 24, 2026 |
| Claude 3.5 SonnetModel by Anthropic | $3.00 | n.a. | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | FeedOct 24, 2026 |
| Claude 3 SonnetModel by Anthropic | $3.00 | n.a. | $15.00 | $33.31 | Batch $1.50 / $7.50 | No change | FeedOct 24, 2026 |
| Nova PremierModel by Amazon | $2.50 | $0.63 | $12.50 | $27.76 | Batch $1.25 / $6.25Flex $1.25 / $6.25Priority $4.38 / $21.88 | No change | FeedOct 24, 2026 |
| Jurassic-2 MidModel by AI21 Labs | $12.50 | n.a. | $12.50 | $96.08 | None published | No change | FeedOct 24, 2026 |
| Mistral LargeModel by Mistral | $4.00 | n.a. | $12.00 | $37.58 | None published | No change | FeedOct 24, 2026 |
| Claude Instant (100K)Model by Anthropic | $3.26 | n.a. | $11.02 | $31.69 | None published | No change | FeedOct 24, 2026 |
| Nova 2.0 ProModel by Amazon | $1.38 | n.a. | $11.00 | $18.79 | Batch $0.69 / $5.50Flex $0.69 / $5.50Priority $2.41 / $19.25Global $1.25 / $10.00 | No change | FeedOct 24, 2026 |
| Claude Sonnet 5.5Model by Anthropic | $2.20 | $0.11 | $11.00 | $24.43 | Global $2.00 / $10.00 | New · Sep 28 | FeedOct 24, 2026 |
| Claude Sonnet 5Model by Anthropic | $2.20 | $0.22 | $11.00 | $24.43 | Global $2.00 / $10.00 | No change | FeedOct 24, 2026 |
| Palmyra X4Model by Writer | $2.50 | n.a. | $10.00 | $25.62 | None published | No change | FeedOct 24, 2026 |
| Jamba 1.5 LargeModel by AI21 Labs | $2.00 | n.a. | $8.00 | $20.50 | None published | No change | FeedOct 24, 2026 |
| xai.grok-4.6Model by xAI | $2.20 | $0.55 | $6.60 | $20.67 | Flex $1.10 / $3.30Priority $3.85 / $11.55Global $2.00 / $6.00 | No change | FeedOct 24, 2026 |
| Grok 4.7Model by xAI | $2.20 | $0.55 | $6.60 | $20.67 | Batch $1.10 / $3.30Flex $1.10 / $3.30Priority $3.85 / $11.55Global $2.00 / $6.00 | New · Sep 29 | FeedOct 24, 2026 |
| Pixtral Large 25.02Model by Mistral | $2.00 | n.a. | $6.00 | $18.79 | None published | No change | FeedOct 24, 2026 |
| Palmyra X5Model by Writer | $0.60 | n.a. | $6.00 | $9.22 | None published | No change | FeedOct 24, 2026 |
| Claude Haiku 4.5Model by Anthropic | $1.10 | $0.11 | $5.50 | $12.21 | Batch $0.55 / $2.75Global $1.00 / $5.00 | No change | FeedOct 24, 2026 |
| DeepSeek R1Model by DeepSeek | $1.35 | n.a. | $5.40 | $13.84 | None published | No change | FeedOct 24, 2026 |
| Nova Pro Latency OptimizedModel by Amazon | $1.00 | n.a. | $4.00 | $10.25 | None published | No change | FeedOct 24, 2026 |
| Claude 3.5 HaikuModel by Anthropic | $0.80 | $0.08 | $4.00 | $8.88 | Batch $0.40 / $2.00 | No change | FeedOct 24, 2026 |
| Llama 3 70BModel by Meta | $2.65 | n.a. | $3.50 | $21.10 | None published | No change | FeedOct 24, 2026 |
| Nova ProModel by Amazon | $0.80 | $0.20 | $3.20 | $8.20 | Batch $0.40 / $1.60Flex $0.40 / $1.60Priority $1.40 / $5.60 | No change | FeedOct 24, 2026 |
| GLM 5Model by Z AI | $1.00 | n.a. | $3.20 | $9.57 | Batch $0.50 / $1.60Flex $0.50 / $1.60Priority $1.75 / $5.60 | No change | FeedOct 24, 2026 |
| Mistral SmallModel by Mistral | $1.00 | n.a. | $3.00 | $9.39 | Batch $0.50 / $1.50 | No change | FeedOct 24, 2026 |
| Kimi K2.5Model by Moonshot AI | $0.60 | n.a. | $3.00 | $6.66 | Batch $0.30 / $1.50Flex $0.30 / $1.50Priority $1.05 / $5.25 | No change | FeedOct 24, 2026 |
| Nova 2.0 OmniModel by Amazon | $0.30 | n.a. | $2.80 | $4.44 | Batch $0.20 / $1.40Flex $0.20 / $1.40Priority $0.60 / $4.80Global $0.30 / $2.50 | No change | FeedOct 24, 2026 |
| Nova 2.0 LiteModel by Amazon | $0.33 | $0.0825 | $2.75 | $4.60 | Batch $0.17 / $1.38Flex $0.17 / $1.38Priority $0.58 / $4.81Global $0.30 / $2.50 | No change | FeedOct 24, 2026 |
| Qwen3 VL 235B A22BModel by Qwen | $0.53 | n.a. | $2.66 | $5.89 | Batch $0.26 / $1.33Flex $0.26 / $1.33Priority $0.93 / $4.66 | No change | FeedOct 24, 2026 |
| Meta Llama 2 Chat 70BModel by Meta | $1.95 | n.a. | $2.56 | $15.51 | None published | No change | FeedOct 24, 2026 |
| xai.grok-4.3Model by xAI | $1.25 | $0.20 | $2.50 | $10.68 | Batch $0.63 / $1.25Flex $0.63 / $1.25Priority $2.19 / $4.38 | No change | FeedOct 24, 2026 |
| Kimi K2 ThinkingModel by Moonshot AI | $0.60 | n.a. | $2.50 | $6.23 | Batch $0.30 / $1.25Flex $0.30 / $1.25Priority $1.05 / $4.38 | No change | FeedOct 24, 2026 |
| Claude InstantModel by Anthropic | $0.80 | n.a. | $2.40 | $7.52 | None published | No change | FeedOct 24, 2026 |
| GLM 4.7Model by Z AI | $0.60 | n.a. | $2.20 | $5.98 | Batch $0.30 / $1.10Flex $0.30 / $1.10Priority $1.05 / $3.85 | No change | FeedOct 24, 2026 |
| DevstralModel by Mistral AI | $0.40 | n.a. | $2.00 | $4.44 | Batch $0.20 / $1.00Flex $0.20 / $1.00Priority $0.70 / $3.50 | No change | FeedOct 24, 2026 |
| Cohere Generate Model - CommandModel by Cohere | $1.00 | n.a. | $2.00 | $8.54 | None published | No change | FeedOct 24, 2026 |
| DeepSeek v3.2Model by DeepSeek | $0.62 | n.a. | $1.85 | $5.82 | Batch $0.31 / $0.93Flex $0.31 / $0.93Priority $1.09 / $3.24 | No change | FeedOct 24, 2026 |
| Qwen3 Coder 480B A35BModel by Qwen | $0.45 | n.a. | $1.80 | $4.61 | Batch $0.23 / $0.90Flex $0.23 / $0.90Priority $0.79 / $3.15 | No change | FeedOct 24, 2026 |
| DeepSeek V3.1Model by DeepSeek | $0.58 | n.a. | $1.68 | $5.40 | Batch $0.29 / $0.84Flex $0.29 / $0.84Priority $1.02 / $2.94 | No change | FeedOct 24, 2026 |
| Mistral Large 3Model by Mistral | $0.50 | n.a. | $1.50 | $4.70 | Batch $0.25 / $0.75Flex $0.25 / $0.75Priority $0.88 / $2.63 | No change | FeedOct 24, 2026 |
| Magistral Small 1.2Model by Mistral | $0.50 | n.a. | $1.50 | $4.70 | Batch $0.25 / $0.75Flex $0.25 / $0.75Priority $0.88 / $2.63 | No change | FeedOct 24, 2026 |
| Cohere Command RModel by Cohere | $0.50 | n.a. | $1.50 | $4.70 | None published | No change | FeedOct 24, 2026 |
| Claude 3 HaikuModel by Anthropic | $0.25 | n.a. | $1.25 | $2.78 | Batch $0.13 / $0.63 | No change | FeedOct 24, 2026 |
| Qwen3 Next 80B A3BModel by Qwen | $0.14 | n.a. | $1.20 | $1.98 | Batch $0.07 / $0.60Flex $0.07 / $0.60Priority $0.26 / $2.10 | No change | FeedOct 24, 2026 |
| Qwen3 Coder NextModel by Qwen | $0.50 | n.a. | $1.20 | $4.44 | Batch $0.25 / $0.60Flex $0.25 / $0.60Priority $0.88 / $2.10 | No change | FeedOct 24, 2026 |
| MiniMax M2.5Model by Minimax AI | $0.30 | n.a. | $1.20 | $3.07 | Batch $0.15 / $0.60Flex $0.15 / $0.60Priority $0.53 / $2.10 | No change | FeedOct 24, 2026 |
| Minimax M2.1Model by Minimax AI | $0.30 | n.a. | $1.20 | $3.07 | Batch $0.15 / $0.60Flex $0.15 / $0.60Priority $0.53 / $2.10 | No change | FeedOct 24, 2026 |
| Minimax M2Model by Minimax AI | $0.30 | n.a. | $1.20 | $3.07 | Batch $0.15 / $0.60Flex $0.15 / $0.60Priority $0.52 / $2.10 | No change | FeedOct 24, 2026 |
| Meta Llama 2 Chat 13BModel by Meta | $0.75 | n.a. | $1.00 | $5.98 | None published | No change | FeedOct 24, 2026 |
| Llama 4 MaverickModel by Meta | $0.24 | n.a. | $0.97 | $2.47 | Batch $0.12 / $0.49 | No change | FeedOct 24, 2026 |
| Llama 3.1 70B Latency OptimizedModel by Meta | $0.90 | n.a. | $0.90 | $6.92 | None published | No change | FeedOct 24, 2026 |
| Qwen3 235B A22B 2507Model by Qwen | $0.22 | n.a. | $0.88 | $2.25 | Batch $0.11 / $0.44Flex $0.11 / $0.44Priority $0.39 / $1.54 | No change | FeedOct 24, 2026 |
| Llama 3.3 70BModel by Meta | $0.72 | n.a. | $0.72 | $5.53 | Batch $0.36 / $0.36 | No change | FeedOct 24, 2026 |
| Llama 3.2 90BModel by Meta | $0.72 | n.a. | $0.72 | $5.53 | None published | No change | FeedOct 24, 2026 |
| Llama 3.1 70BModel by Meta | $0.72 | n.a. | $0.72 | $5.53 | Batch $0.36 / $0.36 | No change | FeedOct 24, 2026 |
| Mixtral 8x7BModel by Mistral | $0.45 | n.a. | $0.70 | $3.67 | None published | No change | FeedOct 24, 2026 |
| Jamba-InstructModel by AI21 Labs | $0.50 | n.a. | $0.70 | $4.01 | None published | No change | FeedOct 24, 2026 |
| Llama 4 Scout 17BModel by Meta | $0.17 | n.a. | $0.66 | $1.73 | Batch $0.085 / $0.33 | No change | FeedOct 24, 2026 |
| NVIDIA Nemotron 3 Super 120B A12BModel by Nvidia | $0.15 | n.a. | $0.65 | $1.58 | Batch $0.075 / $0.33Flex $0.075 / $0.33Priority $0.26 / $1.14 | No change | FeedOct 24, 2026 |
| Writer Palmyra Vision 7BModel by Writer | $0.15 | n.a. | $0.60 | $1.54 | Batch $0.075 / $0.30Flex $0.075 / $0.30Priority $0.26 / $1.05 | No change | FeedOct 24, 2026 |
| Qwen3 Coder 30B A3BModel by Qwen | $0.15 | n.a. | $0.60 | $1.54 | Batch $0.075 / $0.30Flex $0.075 / $0.30Priority $0.26 / $1.05 | No change | FeedOct 24, 2026 |
| Qwen3 32BModel by Qwen | $0.15 | n.a. | $0.60 | $1.54 | Batch $0.075 / $0.30Flex $0.075 / $0.30Priority $0.26 / $1.05 | No change | FeedOct 24, 2026 |
| NVIDIA Nemotron Nano 2 VLModel by Nvidia | $0.20 | n.a. | $0.60 | $1.88 | Batch $0.10 / $0.30Flex $0.10 / $0.30Priority $0.35 / $1.05 | No change | FeedOct 24, 2026 |
| Llama 3 8BModel by Meta | $0.30 | n.a. | $0.60 | $2.56 | None published | No change | FeedOct 24, 2026 |
| gpt-oss-120bModel by OpenAIStandard | $0.15 | n.a. | $0.60 | $1.54 | Batch $0.075 / $0.30Flex $0.075 / $0.30Priority $0.26 / $1.05 | No change | FeedOct 24, 2026 |
| Cohere Generate Model - Command-LightModel by Cohere | $0.30 | n.a. | $0.60 | $2.56 | None published | No change | FeedOct 24, 2026 |
| Claude Haiku 5.5Model by Anthropic | $0.11 | $0.011 | $0.55 | $1.22 | Global $0.10 / $0.50 | New · Oct 7 | FeedOct 24, 2026 |
| Jamba 1.5 MiniModel by AI21 Labs | $0.20 | n.a. | $0.40 | $1.71 | None published | No change | FeedOct 24, 2026 |
| google.gemma-4-31bModel by Google | $0.14 | n.a. | $0.40 | $1.30 | Batch $0.07 / $0.20Flex $0.07 / $0.20Priority $0.25 / $0.70 | No change | FeedOct 24, 2026 |
| google.gemma-4-26b-a4bModel by Google | $0.13 | n.a. | $0.40 | $1.23 | Batch $0.065 / $0.20Flex $0.065 / $0.20Priority $0.23 / $0.70 | No change | FeedOct 24, 2026 |
| GLM 4.7 FlashModel by Z AI | $0.07 | n.a. | $0.40 | $0.82 | Batch $0.035 / $0.20Flex $0.035 / $0.20Priority $0.12 / $0.70 | No change | FeedOct 24, 2026 |
| Gemma 3 27BModel by Google | $0.23 | n.a. | $0.38 | $1.90 | Batch $0.12 / $0.19Flex $0.12 / $0.19Priority $0.40 / $0.67 | No change | FeedOct 24, 2026 |
| gpt-oss-20bModel by OpenAI | $0.07 | n.a. | $0.30 | $0.73 | Batch $0.035 / $0.15Flex $0.035 / $0.15Priority $0.12 / $0.53 | No change | FeedOct 24, 2026 |
| Gemma 3 12BModel by Google | $0.09 | n.a. | $0.29 | $0.86 | Batch $0.05 / $0.15Flex $0.05 / $0.15Priority $0.16 / $0.51 | No change | FeedOct 24, 2026 |
| Nova LiteModel by Amazon | $0.06 | $0.015 | $0.24 | $0.61 | Batch $0.03 / $0.12 | No change | FeedOct 24, 2026 |
| Nemotron Nano 3 30BModel by Nvidia | $0.06 | n.a. | $0.24 | $0.61 | Batch $0.03 / $0.12Flex $0.03 / $0.12Priority $0.11 / $0.42 | No change | FeedOct 24, 2026 |
| NVIDIA Nemotron Nano 2Model by Nvidia | $0.06 | n.a. | $0.23 | $0.61 | Batch $0.03 / $0.12Flex $0.03 / $0.12Priority $0.11 / $0.40 | No change | FeedOct 24, 2026 |
| Llama 3.1 8BModel by Meta | $0.22 | n.a. | $0.22 | $1.69 | Batch $0.11 / $0.11 | No change | FeedOct 24, 2026 |
| Mistral 7BModel by Mistral | $0.15 | n.a. | $0.20 | $1.20 | None published | No change | FeedOct 24, 2026 |
| Ministral 14B 3.0Model by Mistral | $0.20 | n.a. | $0.20 | $1.54 | Batch $0.10 / $0.10Flex $0.10 / $0.10Priority $0.35 / $0.35 | No change | FeedOct 24, 2026 |
| Llama 3.2 11BModel by Meta | $0.16 | n.a. | $0.16 | $1.23 | None published | No change | FeedOct 24, 2026 |
| Ministral 8B 3.0Model by Mistral | $0.15 | n.a. | $0.15 | $1.15 | Batch $0.07 / $0.07Flex $0.07 / $0.07Priority $0.26 / $0.26 | No change | FeedOct 24, 2026 |
| Llama 3.2 3BModel by Meta | $0.15 | n.a. | $0.15 | $1.15 | None published | No change | FeedOct 24, 2026 |
| Nova MicroModel by Amazon | $0.035 | $0.0088 | $0.14 | $0.36 | Batch $0.0175 / $0.07 | No change | FeedOct 24, 2026 |
| Ministral 3B 3.0Model by Mistral | $0.10 | n.a. | $0.10 | $0.77 | Batch $0.05 / $0.05Flex $0.05 / $0.05Priority $0.17 / $0.17 | No change | FeedOct 24, 2026 |
| Llama 3.2 1BModel by Meta | $0.10 | n.a. | $0.10 | $0.77 | None published | No change | FeedOct 24, 2026 |
| google.gemma-4-e2bModel by Google | $0.04 | n.a. | $0.08 | $0.34 | Batch $0.02 / $0.04Flex $0.02 / $0.04Priority $0.07 / $0.14 | No change | FeedOct 24, 2026 |
| Gemma 3 4BModel by Google | $0.04 | n.a. | $0.08 | $0.34 | Batch $0.02 / $0.04Flex $0.02 / $0.04Priority $0.07 / $0.14 | No change | FeedOct 24, 2026 |

Revenue per GPU-hour is SAVRN's calculation, not a price any seller charges: the 6.83 million input and 0.85 million output tokens one Vera Rubin GPU handles in an hour in CoreWeave's measured test, billed at this model's list prices. It puts a token price beside a GPU rental price; it says nothing about how fast this model actually runs.

Committed capacity

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| AWS Bedrock | Provisioned Throughput (Cohere Command)Sold per Model Unit-hour | $49.50 per Model Unit-hour$49.50 per Model Unit-hour with no commitment; $39.60 with a 1-month commitment; $23.77 with a 6-month commitment | no-commit / 1-mo / 6-mo (−52%)Anthropic Claude PT: "reach out to your account team" | FeedOct 24, 2026[aws.amazon.com ↗](https://aws.amazon.com/bedrock/pricing/) |
| AWS Bedrock | Provisioned Throughput (Nova family)Sold per Model Unit-hour | $60.50 per Model Unit-hour$60.50 per Model Unit-hour with no commitment; $55.00 with a 1-month commitment; $30.25 with a 6-month commitment | 6-month commit −50%Model Unit = fixed tokens/minute; worked example 2 MU Titan Express = $27,379/month | FeedOct 24, 2026[aws.amazon.com ↗](https://aws.amazon.com/bedrock/pricing/) |

Dedicated capacity

## Committed capacity from hosts and GPU clouds

Model capacity sold by the hour, the month or the contract instead of by the token.

| Seller | Product | Published price | Term and capacity | Checked |
| --- | --- | --- | --- | --- |
| Cohere | Model Vault dedicatedSold per Hour or month | Embed 4 small $4/hr or $2,500/mo | MonthlyDedicated model instance | ResearchOct 21, 2026[cohere.com ↗](https://cohere.com/pricing) |
| Together AI | dedicated endpointsSold per GPU-hour | H100 $3.99; B200 $8.99 | Reserved clusters H100 $1.99–3.19 (to −50%)Dedicated capacity sold by GPU-hour, tokens uncapped | ResearchOct 21, 2026[together.ai ↗](https://www.together.ai/pricing) |
| SF Compute | forward marketSold per GPU-hour | Realized $4.29; reserved $3.00; average $1.93 | One node-hour to multi-yearTradeable forward contracts on H100/B300/GB300 | ResearchOct 21, 2026[sfcompute.com ↗](https://sfcompute.com/) |
| Fireworks | on-demand dedicated (from 1 Sep 2026)Sold per GPU-hour | H100 $8; B200 $13; B300 $15; GB300 $20 | None; region-restricted 1.5×Dedicated capacity by GPU-hour | ResearchOct 21, 2026[fireworks.ai ↗](https://fireworks.ai/pricing) |

The SAVRN reference set

## Seven references, one system

Capital, delay, policy, water, grid, permits and scarcity are one system seen from seven sides. Each reference is primary-sourced, refreshed on a schedule, and links into the others, so a project can be followed from the money committed to the docket that stopped it.

Capital Atlas

Who is funding AI data centers: capital providers, named sites, financing transactions, and the blocked-and-delayed record, state by state and company by company.

[Open the atlas →](https://savrn.com/data-center-capital-atlas)

Delay Watchlist

Project-level slippage: named projects, jurisdictional actions, utility and federal dockets, SEC disclosures, and the interconnection queue.

Open the watchlist →

Moratorium Tracker

The policy layer: every state measure and local moratorium mapped and cited, with one page per state.

Open the tracker →

Water Tracker

Every federal, state, local and interstate water measure that touches a data center, plus a fifty-state water ledger.

Open the water tracker →

Grid Operator Watchlist

ERCOT, PJM, MISO, CAISO, SPP, NYISO, ISO-NE and FERC: the large-load rules, filings and queue positions that decide whether a site can energize.

Open the watchlist →

Permits and Power Development

Data center and substation permits, large loads in the interconnection queues, and the power plants planned to serve them, refreshed hourly from public records.

Open the permits tracker →

SAVRN Scarcity Tracker

Eleven primary-sourced indicators of AI infrastructure scarcity, one composite index, read against the token index, SAVRN's own cost basis and the frontier ceiling.

Open the scarcity tracker →

Have a specific address? Site Intelligence reads all seven references against one location, county, utility and grid operator included, and cites every finding.

[Check an address →](https://savrn.com/site-intelligence)