What Is Claude Mythos and What Does It Cost?
Claude Mythos 5.1 costs $10 per million input tokens and $50 per million output tokens. That is exactly double Claude Opus 5, which is $5 and $25 for the same volumes.
Mythos sits at the top of Anthropic's published line, and availability is limited rather than open to every account. Claude Fable 5.1, released in September 2026 and described by Anthropic as Mythos-class, carries identical $10 and $50 pricing.
Prices here were checked on 16 September 2026 against Anthropic's and OpenAI's own pricing documentation. AI Perks tracks $7.7M in credits across 194 companies that offset these bills.
Claude Mythos 5.1 vs the Rest of the Claude Line
Anthropic's entire current line runs a clean 5x output multiple, so choosing a tier is a single decision about price per token rather than a decision about token economics.
| Model | Input per 1M | Output per 1M | 1M in + 1M out | Output multiple |
|---|---|---|---|---|
| Claude Fable 5.1 | $10.00 | $50.00 | $60.00 | 5x |
| Claude Mythos 5.1 | $10.00 | $50.00 | $60.00 | 5x |
| Claude Opus 5 | $5.00 | $25.00 | $30.00 | 5x |
| Claude Sonnet 5 | $2.00 | $10.00 | $12.00 | 5x |
| Claude Haiku 4.5 | $1.00 | $5.00 | $6.00 | 5x |
The combined column is derived arithmetic, not a published figure.
Read down the input column and the ladder is exact: Mythos is 2x Opus 5, 5x Sonnet 5, and 10x Haiku 4.5 on both input and output. Claude Opus 5 launched on 24 July 2026 at the same price as Opus 4.8, so the step up to Mythos is the only price move Anthropic has made at the top of the line.
Credit programs are what make the top two tiers survivable for a small team, and AI Perks lists which providers run them.
How Claude Mythos Compares Across Vendors
Claude Mythos 5.1, Claude Fable 5.1 and OpenAI's gpt-6-astra all list at $10 input and $50 output. The 2026 flagship price has converged to the same number at both vendors.
| Model | Input per 1M | Output per 1M | Output multiple |
|---|---|---|---|
| Claude Mythos 5.1 | $10.00 | $50.00 | 5.0x |
| Claude Fable 5.1 | $10.00 | $50.00 | 5.0x |
| OpenAI gpt-6-astra | $10.00 | $50.00 | 5.0x |
| OpenAI gpt-6-astra above 200k context | $20.00 | $75.00 | 3.75x |
| OpenAI gpt-5.5 | $5.00 | $30.00 | 6.0x |
| Claude Opus 5 | $5.00 | $25.00 | 5.0x |
| OpenAI gpt-5.6-sol | $4.00 | $20.00 | 5.0x |
| OpenAI gpt-5.4 | $2.50 | $15.00 | 6.0x |
| Claude Sonnet 5 | $2.00 | $10.00 | 5.0x |
| OpenAI gpt-5.6-terra | $2.00 | $12.00 | 6.0x |
| Meta Muse Spark 1.3 | $1.25 | $4.25 | 3.4x |
| Claude Haiku 4.5 | $1.00 | $5.00 | 5.0x |
| OpenAI gpt-5.4-mini | $0.75 | $4.50 | 6.0x |
| OpenAI gpt-5.6-luna | $0.20 | $1.20 | 6.0x |
| OpenAI gpt-5.4-nano | $0.20 | $1.25 | 6.25x |
| Meta Muse Spark Contributor | $0.10 | $0.20 | 2.0x |
Two things fall out of the output multiple column. OpenAI's mid and low tiers charge 6x for output against Anthropic's flat 5x, and Meta Muse Spark charges only 3.4x, the lowest ratio of any full price model here.
OpenAI is the only vendor in this table publishing a long context surcharge: above 200,000 tokens, gpt-6-astra moves to $20 and $75. Anthropic has not published an equivalent surcharge for Mythos.
Google Gemini prices are deliberately absent because they were not verified in this research. xAI's Grok token prices are also absent: published sources conflict badly enough that the only honest answer is that pricing varies by model and belongs on xAI's own console.
What Claude Mythos Costs Per Month at Real Volume
At a workload of 50 million input tokens and 10 million output tokens per month, Claude Mythos 5.1 runs an estimated $1,000. The same workload on Claude Haiku 4.5 is an estimated $100.
| Model | Input cost | Output cost | Estimated monthly |
|---|---|---|---|
| Claude Mythos 5.1 | $500 | $500 | $1,000 |
| Claude Fable 5.1 | $500 | $500 | $1,000 |
| OpenAI gpt-6-astra | $500 | $500 | $1,000 |
| Claude Opus 5 | $250 | $250 | $500 |
| OpenAI gpt-5.6-sol | $200 | $200 | $400 |
| OpenAI gpt-5.6-terra | $100 | $120 | $220 |
| Claude Sonnet 5 | $100 | $100 | $200 |
| Meta Muse Spark 1.3 | $62.50 | $42.50 | $105 |
| Claude Haiku 4.5 | $50 | $50 | $100 |
| OpenAI gpt-5.6-luna | $10 | $12 | $22 |
These are estimates derived from list prices at a stated volume, with no discounts, caching or batch pricing applied.
The spread is the useful part. Top tier to cheapest current tier is 45x on identical traffic, which is the difference between a $12,000 and a $264 annual API line. A credit grant that covers a year of Haiku barely dents a year of Mythos, which is why AI Perks is worth checking before you pick a default model.
The Sonnet 5 Price Rise That Never Happened
Anthropic had scheduled Claude Sonnet 5 to rise from $2 and $10 to $3 and $15 on 1 September 2026, then cancelled the increase. The $2 and $10 price is now standard, not introductory.
Anthropic's pricing page states it plainly: "The previously scheduled increase to $3/$15 per million input/output tokens on September 1, 2026 will not occur."
This is the most under-reported number in the current Claude line. At the 50M and 10M workload above, the cancelled rise is worth an estimated $100 per month, a 33% saving, or roughly $1,200 a year on one model choice.
It also changes the Mythos maths. Against $2 and $10, Mythos 5.1 is a 5x step. Against the scheduled price it would have been 3.3x, so anyone who budgeted in August used the wrong ratio.
Which Claude Tier Fits Your Workload
Pick by output volume, because output is priced at 5x input on every Anthropic tier.
Claude Mythos 5.1 or Claude Fable 5.1 at $10 and $50. For work where a single wrong answer costs more than the token bill. Mythos availability is limited, so Fable 5.1 is the practically obtainable model at that price point.
Claude Opus 5 at $5 and $25. Half the flagship price with no increase since Opus 4.8. The default for serious agent and coding work that runs continuously.
Claude Sonnet 5 at $2 and $10. The best value in the line now that the increase is cancelled. High volume production traffic belongs here.
Claude Haiku 4.5 at $1 and $5. Classification, extraction, routing and anything you run millions of times.
For output-heavy workloads, invert the reading: at 10M input and 50M output, Mythos is an estimated $2,600 a month against $520 for Sonnet 5 and $225 for Meta Muse Spark 1.3. Note that Meta's cheapest Contributor tier at $0.10 and $0.20 trains on your prompts and completions, and Meta runs no startup credit program for Muse. AI Perks covers the providers that do.
What Has Not Been Published, and What Is Shutting Down
Anthropic has not published benchmark scores, context window sizes, or a long context surcharge for Claude Mythos 5.1 in the material this page was verified against. Anything you read claiming those numbers is inference, not documentation.
The unknowns for Mythos: no capability ranking against Opus 5, no parameter or architecture detail, no published rate limits. The one hard fact beyond price is limited availability.
On the OpenAI side there is a dated deadline worth putting in a calendar. The legacy line shuts down on 10 December 2026:
| Legacy model | Input per 1M | Output per 1M | Status |
|---|---|---|---|
| gpt-5 | $1.25 | $10.00 | API shutdown 10 Dec 2026 |
| gpt-5-mini | $0.25 | $2.00 | API shutdown 10 Dec 2026 |
| gpt-5-nano | $0.05 | $0.40 | API shutdown 10 Dec 2026 |
Do not start new builds on these. The Sol, Terra and Luna names replaced the older Instant, Thinking, mini and nano scheme in July 2026, and gpt-5.6-luna at $0.20 and $1.20 is the current cheap tier.
Frequently Asked Questions
How much does Claude Mythos 5.1 cost per million tokens?
Claude Mythos 5.1 costs $10 per million input tokens and $50 per million output tokens. That is exactly double Claude Opus 5 at $5 and $25, and five times Claude Sonnet 5 at $2 and $10. Credits that offset those bills are tracked at getaiperks.com.
Is Claude Mythos more expensive than GPT-6?
No, they list identically. Claude Mythos 5.1, Claude Fable 5.1 and OpenAI's gpt-6-astra all price at $10 input and $50 output per million tokens. The difference appears above 200,000 tokens, where gpt-6-astra moves to $20 and $75. Anthropic has published no equivalent surcharge.
Can I actually get access to Claude Mythos?
Anthropic lists Claude Mythos 5.1 as limited availability. Claude Fable 5.1, released September 2026 and described as Mythos-class, sits at the same $10 and $50 pricing and is the practical route to that tier. Free and discounted API credits across 194 companies are listed at getaiperks.com.
Did Claude Sonnet 5 get more expensive in September 2026?
No. Anthropic scheduled an increase to $3 and $15 for 1 September 2026, then cancelled it. The page states the increase "will not occur", so $2 and $10 is now the standard price. That is an estimated $1,200 a year saved at moderate volume.
What is the cheapest way to run production AI traffic in 2026?
By list price, OpenAI's gpt-5.6-luna at $0.20 and $1.20, then Meta Muse Spark Contributor at $0.10 and $0.20, which trains on your prompts. On Anthropic, Claude Haiku 4.5 at $1 and $5. Credits change the answer entirely, and AI Perks tracks $7.7M of them.
How does Claude Mythos compare to Grok on price?
xAI's published token prices conflict badly across sources, so no figure is printed here. Check current rates in xAI's own console before budgeting. For Anthropic, the verified line runs $1 and $5 for Haiku 4.5 up to $10 and $50 for Mythos 5.1.
Pick the tier your workload needs. Let someone else cover the invoice.