@ArtificialAnlys: Claude Fable 5 cost ~$6.2K to ...
@ArtificialAnlys
32 views
Jun 18, 2026
Advertisement
1
Claude Fable 5 cost ~$6.2K to run the Artificial Analysis Intelligence Index benchmarks - the most expensive model we have ever benchmarked 🧵
Key takeaways:
➤ Intelligence Index: 60, ahead of Claude Opus 4.8 (56) and GPT-5.5 (55)
➤ Cost to run the Intelligence Index: $6.2K, 1.7× the next-highest model (Opus 4.8, $3.7K) and 2.2× GPT-5.5 (xhigh, $2.9K)
➤ List price: $10/$50 per 1M input/output tokens, 2× Opus 4.8. Among 2026 releases, only OpenAI's special Pro tier (GPT-5.5 Pro, $30/$180) is priced higher
➤ Cache pricing, which is particularly important for long agentic coding sessions, doubled too: $1/M cache reads, and $12.50/M cache writes vs $0.50/$6.25 for Opus 4.8
➤ The top 3 most-expensive models to run the Intelligence Index are now all Claude models
Key takeaways:
➤ Intelligence Index: 60, ahead of Claude Opus 4.8 (56) and GPT-5.5 (55)
➤ Cost to run the Intelligence Index: $6.2K, 1.7× the next-highest model (Opus 4.8, $3.7K) and 2.2× GPT-5.5 (xhigh, $2.9K)
➤ List price: $10/$50 per 1M input/output tokens, 2× Opus 4.8. Among 2026 releases, only OpenAI's special Pro tier (GPT-5.5 Pro, $30/$180) is priced higher
➤ Cache pricing, which is particularly important for long agentic coding sessions, doubled too: $1/M cache reads, and $12.50/M cache writes vs $0.50/$6.25 for Opus 4.8
➤ The top 3 most-expensive models to run the Intelligence Index are now all Claude models
2
Where does the $6,228 come from? The token counts, plus our v4.1 cost methodology:
Fable 5 consumed 676M tokens to run our Intelligence Index: 590M input, 76.5M reasoning, and 9.8M answer tokens. At Fable 5's, without considering cached inputs would be ~$10,215, but v4.1 prices cached tokens at cache rates and credits fallback tasks to Opus 4.8, bringing the total cost down to $6,228.
Highlights to consider:
➤ Cache is the biggest lever: most of Fable 5's input is cached, billed at $1/M (reads) and $12.50/M (writes) rather than $10/M, cutting input cost from ~$5,900 to ~$2,050
➤ Automatic fallback: ~8% of tasks were served by Claude Opus 4.8 at $5/$25, trimming a further ~$275 off
➤ The record cost is not driven by verbosity: Fable 5 generated ~26M fewer output tokens than Opus 4.8. Cost increase is mainly driven by the 2× token price increase.
Fable 5 consumed 676M tokens to run our Intelligence Index: 590M input, 76.5M reasoning, and 9.8M answer tokens. At Fable 5's, without considering cached inputs would be ~$10,215, but v4.1 prices cached tokens at cache rates and credits fallback tasks to Opus 4.8, bringing the total cost down to $6,228.
Highlights to consider:
➤ Cache is the biggest lever: most of Fable 5's input is cached, billed at $1/M (reads) and $12.50/M (writes) rather than $10/M, cutting input cost from ~$5,900 to ~$2,050
➤ Automatic fallback: ~8% of tasks were served by Claude Opus 4.8 at $5/$25, trimming a further ~$275 off
➤ The record cost is not driven by verbosity: Fable 5 generated ~26M fewer output tokens than Opus 4.8. Cost increase is mainly driven by the 2× token price increase.
3
Claude now has four pricing tiers, and the new top rung is priced at exactly 2× the old one.
Fable 5 is the first Claude tier to launch above Opus, doubling flagship pricing across every token type. Anthropic's lineup now spans 10× from bottom to top (USD per 1M input/output tokens):
➤ Haiku 4.5 → $1 / $5
➤ Sonnet 4.6 → $3 / $15
➤ Opus 4.8 → $5 / $25
➤ Fable 5 → $10 / $50
It is also the first time Anthropic has raised its flagship price: $15/$75 held from Claude 3 Opus through Opus 4.1, was cut to $5/$25 with Opus 4.5, and stayed there through Opus 4.8 until now.
Fable 5 is the first Claude tier to launch above Opus, doubling flagship pricing across every token type. Anthropic's lineup now spans 10× from bottom to top (USD per 1M input/output tokens):
➤ Haiku 4.5 → $1 / $5
➤ Sonnet 4.6 → $3 / $15
➤ Opus 4.8 → $5 / $25
➤ Fable 5 → $10 / $50
It is also the first time Anthropic has raised its flagship price: $15/$75 held from Claude 3 Opus through Opus 4.1, was cut to $5/$25 with Opus 4.5, and stayed there through Opus 4.8 until now.
4
Methodology note: these figures use v4.1 of the Intelligence Index, which now prices cached tokens from agentic evaluations at cache read/write rates rather than full input price, better reflecting real cost on cache-heavy evals.
Where a model routes requests to a fallback, our cost-to-run figures also price each token at the rates of the model that actually served it, and Fable 5 appears on our charts as "Claude Fable 5 (with fallback)".
In the future, we will also break out intelligence scoring for models with fallback or refusal behavior, so it is visible how much of a score is earned by the headline model itself.
Judge whether the intelligence gains justify the premium: see Intelligence vs Cost 👇
artificialanalysis.ai/?cost=intellig…
Where a model routes requests to a fallback, our cost-to-run figures also price each token at the rates of the model that actually served it, and Fable 5 appears on our charts as "Claude Fable 5 (with fallback)".
In the future, we will also break out intelligence scoring for models with fallback or refusal behavior, so it is visible how much of a score is earned by the headline model itself.
Judge whether the intelligence gains justify the premium: see Intelligence vs Cost 👇
artificialanalysis.ai/?cost=intellig…