Anthropic holds the cheapest of 6 offers for claude-fable-5.2
Anthropic charges an Input price of $8.00 and an Output price of $40.00 per million tokens for claude-fable-5.2, checked on November 12, 2026. Cache reads of claude-fable-5.2 cost $0.20 per million tokens at Anthropic. Our rule designates Anthropic as the cheapest of 6 sellers, with a mixed 3:1 price of $16.00. The comparison of these 6 sellers is not under verification.
Per million tokens, claude-fable-5.2 at Anthropic, checked on November 12, 2026.
| Seller | Input, $/M | Output, $/M | Mixed 3:1, $/M |
|---|---|---|---|
| Anthropic | 8.00 | 40.00 | 16.00 |
| Vercel AI Gateway | 8.00 | 40.00 | 16.00 |
| ZenMux | 8.00 | 40.00 | 16.00 |
| Amazon Bedrock | 8.00 | 40.00 | 16.00 |
| Azure | 8.00 | 40.00 | 16.00 |
| NanoGPT | 10.00 | 50.00 | 20.00 |
claude-fable-5.2 carries one grid at Anthropic, Vercel AI Gateway, ZenMux, Amazon Bedrock and Azure, and a higher grid at NanoGPT.Vercel AI Gateway, ZenMux, Amazon Bedrock and Azure match Anthropic's grid for claude-fable-5.2
Vercel AI Gateway, ZenMux, Amazon Bedrock and Azure each list claude-fable-5.2 at Input $8.00, Output $40.00 and cache reads $0.20 per million tokens. Their prices were checked on November 12, 2026. At each of these sellers, the mixed 3:1 price of claude-fable-5.2 is $16.00.
A team already calling claude-fable-5.2 through Anthropic, Vercel AI Gateway, ZenMux, Amazon Bedrock or Azure pays the same listed grid at each of them. For that team, the listed prices give no reason to move between these sellers.
NanoGPT lists claude-fable-5.2 at Input $10.00, Output $50.00 and cache reads $0.25 per million tokens, checked on November 12, 2026. The mixed 3:1 price of claude-fable-5.2 at NanoGPT is $20.00. The sources do not say why the NanoGPT grid differs.
claude-fable-5.2 whose mixed price stands apart from the others.claude-fable-5.2 accepts up to 1,000,000 tokens of context at Anthropic
Tokens, claude-fable-5.2, the full window carried by the Anthropic offer.
The largest context window recorded for claude-fable-5.2 is 1,000,000 tokens, and the Anthropic offer carries that full window. The index holds context data for claude-fable-5.2 at 7 sellers, while its price comparison covers 6 sellers. A reader who needs the full window can start from the Anthropic offer and check the window on the page of any other seller.
Cache-heavy and output-heavy workloads on claude-fable-5.2 weigh on different lines
An agent that resends a long, stable prompt to claude-fable-5.2 at Anthropic pays cache reads at $0.20 per million tokens, against Input at $8.00. A workload that writes long answers pays Output for claude-fable-5.2 at $40.00 per million tokens at Anthropic, the largest line of its grid. The mixed 3:1 price assumes three Input tokens for each Output token, so it understates the bill of an output-heavy job.
At NanoGPT, Output for claude-fable-5.2 is $50.00 per million tokens and cache reads are $0.25 per million tokens. For long answers and heavy caching alike, the NanoGPT grid is the one to price separately.
Before routing traffic to claude-fable-5.2, confirm the seller's grid and window
- Read the Output line first
If your answers are long, compare Output for
claude-fable-5.2: $40.00 per million tokens at Anthropic, $50.00 at NanoGPT. Success: you know which line drives your bill. - Measure your cache read share
Log how many prompt tokens are served as cache reads; for
claude-fable-5.2at Anthropic they cost $0.20 per million tokens. Success: the share is known before you pick a seller. - Confirm the context window
Check that the chosen seller exposes the 1,000,000-token window that the Anthropic offer of
claude-fable-5.2carries. Success: your longest prompt fits without truncation. - Recheck the seller's page
The prices of
claude-fable-5.2here were checked on November 12, 2026. Success: the grid on the seller's page matches the table on the day you switch.