This is the real Anthropic decision in September 2026, because both models are strong and the price gap is 2x. Claude Fable 5.1 is the capability ceiling: 52.6% on Terminal-Bench-Science 0.1 against Opus 5's 29.0%, and 55.8% on Terminal-Bench 4.0 against 52.3%. Claude Opus 5 costs $5/$25 per 1M tokens against Fable 5.1's $10/$50, and it holds the published SWE-bench Verified lead at roughly 96%. Read the two benchmarks that separate them: on agentic coding the gap is 3.5 points, which rarely justifies double the price; on agentic scientific research it is 23.6 points, which absolutely does. Fable 5.1's 75% cache-read cut ($0.25 per 1M) narrows the real-world gap further on agent loops that replay a large cached prompt every turn.
AnthropicPremium
Claude Fable 5.1
New frontier leader — better than Fable 5 on every published benchmark, and cheaper to run.
Winner
VS
AnthropicPremium
Claude Opus 5
Best premium model for agentic coding — near-Fable 5 quality at half the price.
At a glance
Claude Fable 5.1
Claude Opus 5
Input cost / 1M tokens
$$10.00/1M
$$5.00/1M
Output cost / 1M tokens
$$50.00/1M
$$25.00/1M
Context window
1M tokens
1M tokens
Speed
Deliberate
Deliberate
Price tier
Premium
Premium
Benchmarks
SWE-bench (coding)
—
96%
Arena Elo
—
—
MMLU
—
—
How they compare
Which model wins for each use case — and why.
Frontier ceilingClaude Fable 5.1 wins
Fable 5.1 leads Terminal-Bench-Science 0.1 by 23.6 points (52.6% vs 29.0%) and AutomationBench by 4.5 (31.4% vs 26.9%).
Price / valueClaude Opus 5 wins
Opus 5 is $5/$25 against $10/$50 — half the base price for a model that trails by only 3.5 points on Terminal-Bench 4.0 agentic coding.
Repository-level engineeringClaude Opus 5 wins
Opus 5 holds the published SWE-bench Verified lead at about 96%. Anthropic published no SWE-bench figure for Fable 5.1, so on that benchmark Opus 5 is the one with a score on the record.
Agentic codingClaude Fable 5.1 wins
Terminal-Bench 4.0: 55.8% for Fable 5.1 against 52.3% for Opus 5 — a real but narrow lead.
Computer useClaude Fable 5.1 wins
OSWorld 2.0 strict: 41.7% for Fable 5.1 against 39.6% for Opus 5, with the same ordering on the partial-credit variant (77.9% vs 75.4%).
Context & output limitsTie
Both provide a 1M-token context window and 128K max output at standard per-token rates.
Which should you pick?
Pick Claude Fable 5.1 if…
Your work is agentic scientific research or long autonomous loops, where the gap is 20+ points
Your agents are cache-heavy, so the $0.25/1M cache-read rate absorbs much of the 2x base price
Capability is the bottleneck and the API bill is not
You want the highest published agentic-coding and computer-use scores available today
For most workflows, Claude Fable 5.1 is the stronger choice.
The new frontier leader, and the rare upgrade that costs less than the model it replaces. Base $10/$50 pricing did not move, but a 75% cut to cache reads makes typical workloads ~25% cheaper than Fable 5 and heavily agentic ones ~45% cheaper — while more than doubling Fable 5 on agentic science and beating Opus 5 on agentic coding. If you were already on Fable 5, switching is a straight win. If you were on Opus 5 for the price, it still costs 2x on uncached tokens.
The case for each model
What each one is genuinely good at, where it falls down, and when we would steer you away from it.
Anthropic's September 1, 2026 frontier release and the new capability ceiling for coding, agents, and scientific work. Base pricing is unchanged at $10/$50, but cache reads dropped 75% to $0.25/1M — roughly 25% cheaper on typical workloads and up to 45% cheaper on agentic ones. 1M context, 128K output, adaptive thinking always on.
Input
$10.00/1M
Output
$50.00/1M
Context
1M tokens
Speed
Deliberate
What people actually use it for
Agentic scientific research — 52.6% on Terminal-Bench-Science 0.1, more than double Fable 5's 24.7%
Long-running autonomous coding agents that plan, run, and debug across a whole repository
Cache-heavy agent loops where the 75% cache-read cut ($1.00 → $0.25 per 1M) is the real saving
Where it wins
52.6% Terminal-Bench-Science 0.1 — 2.1x Fable 5 (24.7%), well clear of Opus 5 (29.0%) and GPT-5.6 Sol (22.4%)
55.8% Terminal-Bench 4.0 agentic coding, ahead of Opus 5 (52.3%) and Fable 5 (42.0%)
Cache reads cut 75% to $0.25/1M — ~25% cheaper for typical use, ~45% for heavily agentic work
Independent Vals AI evaluation ranks it #1 of 51 on the Vals Index, #1 on LiveCodeBench (90.5%) and MMLU Pro (92.4%)
1M-token context and 128K max output at standard rates, with adaptive thinking always enabled
Where it falls down
Base rates are still $10/$50 per 1M — double Claude Opus 5 for anything that is not cache-heavy
Anthropic published no SWE-bench Verified or SWE-bench Pro figure for 5.1 at launch
On Claude Pro it only runs on pay-as-you-go credits; Max includes it up to 50% of weekly limits
Deliberate, high-latency profile — the wrong choice for interactive, latency-bound apps
Skip it if
You need low latency, or your workload is uncached and cost-sensitive — Claude Opus 5 is half the base price and within a few points on agentic coding.
Our verdict
The new frontier leader, and the rare upgrade that costs less than the model it replaces. Base $10/$50 pricing did not move, but a 75% cut to cache reads makes typical workloads ~25% cheaper than Fable 5 and heavily agentic ones ~45% cheaper — while more than doubling Fable 5 on agentic science and beating Opus 5 on agentic coding. If you were already on Fable 5, switching is a straight win. If you were on Opus 5 for the price, it still costs 2x on uncached tokens.
Released September 1, 2026 alongside Claude Mythos 5.1, the first update to the Mythos-class line since Fable 5 on June 9. API ID claude-fable-5-1; generally available on the Claude API, Amazon Bedrock, Google Cloud, and Microsoft Foundry. Published launch numbers (Fable 5.1 / Fable 5 / Opus 5 / GPT-5.6 Sol): Terminal-Bench-Science 0.1 52.6 / 24.7 / 29.0 / 22.4; Terminal-Bench 4.0 55.8 / 42.0 / 52.3 / 37.3; CursorBench 3.2.0 73.4 / 70.5 / 70.0 / 67.2; AutomationBench 31.4 / 17.1 / 26.9 / 19.6; OSWorld 2.0 strict 41.7 / 36.1 / 39.6; Humanity's Last Exam (no tools) 60.9 / 57.8 / 56.6; GDPval-AA v2 1853 / 1723 / 1824 / 1711. GDPval-AA v2 is rescaled from the v1 numbers quoted on the Fable 5 page and is not directly comparable to them.
Anthropic's flagship-tier Opus that comes close to Claude Fable 5's frontier intelligence at half the price — the new default for complex agentic coding and enterprise agents.
Input
$5.00/1M
Output
$25.00/1M
Context
1M tokens
Speed
Deliberate
What people actually use it for
Long-running repository-level engineering with the top SWE-bench Verified score (~96%)
Autonomous computer-use agents — beats Fable 5's best OSWorld 2.0 result at one-third the cost
Enterprise automation where Opus 4.8 quality was the bar and price capped Fable 5 adoption
Where it wins
Leads SWE-bench Verified at ~96% — the top public coding score as of August 2026
89.1% on Terminal-Bench 2.1 and state-of-the-art on Frontier-Bench v0.1, more than doubling Opus 4.8's result
Within 0.5% of Claude Fable 5 on CursorBench 3.2 at half the cost per task
Where it falls down
Still short of Claude Fable 5's absolute frontier ceiling on the hardest long-running agent tasks
Fast mode (research preview) doubles pricing to $10/$50, erasing the cost advantage over Fable 5
Skip it if
You need fast interactive latency (Sonnet 5 is quicker and one-fifth the price) or absolute frontier ceiling (Fable 5).
Our verdict
The new premium coding default. Opus 5 takes the SWE-bench Verified lead from Opus 4.8 at the same $5/$25 price, and gets within half a point of Fable 5 on agentic coding at half the cost. Choose Fable 5 only when absolute frontier intelligence matters more than budget.
Released July 24, 2026 at Opus 4.8's exact pricing. 1M context at standard rates, 128K max output. Anthropic's alignment audit calls it their most aligned model to date. Default model on Claude Max plans.
Frequently asked questions
Is Claude Fable 5.1 worth double the price of Opus 5?
It depends on the task. On agentic coding the gap is 3.5 points on Terminal-Bench 4.0 — usually not worth 2x. On agentic scientific research it is 23.6 points (52.6% vs 29.0%), which is a different class of result. Cache-heavy workloads also shrink the real gap, because Fable 5.1 reads cache at $0.25 per 1M.
Which is better for coding, Fable 5.1 or Opus 5?
Fable 5.1 leads agentic coding on Terminal-Bench 4.0 (55.8% vs 52.3%) and CursorBench 3.2.0 (73.4% vs 70.0%). Opus 5 holds the published SWE-bench Verified lead at about 96% and Anthropic published no SWE-bench figure for 5.1. For repo-level issue fixing at volume, Opus 5 is the better value; for long autonomous agent runs, Fable 5.1.
How much do Fable 5.1 and Opus 5 cost?
Fable 5.1 is $10 per 1M input tokens and $50 per 1M output, with cache reads at $0.25 per 1M. Opus 5 is $5 and $25. On a 10M-input / 2M-output month that is $200 against $100 before any caching discount.
Are Fable 5.1 and Opus 5 the same context window?
Yes. Both offer a 1M-token context window and 128K max output tokens per request, billed at standard per-token rates.
Which model should most teams default to?
Opus 5 for everyday and high-volume work, Fable 5.1 for the hardest agentic and research tasks. Routing between them by task type is cheaper than standardising on either one.