Claude Fable 5
AnthropicJun 9, 2026N/A
Anthropic's new premium line, separate from Opus/Sonnet/Haiku. Highest peak-effort quality in Anthropic's range: #1 on the LMArena text board (~1525 Elo) and the SWE-bench Pro leader (~80%). Max-effort runs spend far more tokens than Opus 5.
- Status
- Generally available
- License
- Proprietary API
- Context window
- 1M
- Max output
- 128K
- Input · $/M
- $8
- Output · $/M
- $40
- Cached input · $/M
- $0.8
- Blended · $/M
- $16.00
- Parameters
- Undisclosed
- Inputs
- Text + images
Benchmarks
Composite 80.0· mean of published scores
Reasoning & knowledge
GPQA Diamond
no public score
MMLU-Pro
no public score
Humanity's Last Exam
no public score
ARC-AGI-2
no public score
Coding
SWE-bench Verified
no public score
SWE-bench ProLeader
80%
LiveCodeBench
no public score
Agents & computer use
Terminal-Bench 2
no public score
OSWorld 2.0
no public score
BrowseComp
no public score
Math
AIME 2026
no public score
Known strengths
- #1 LMArena text Elo
- SWE-bench Pro leader (80%+)
- DeepSWE 66–70% at max effort
- Long-horizon agency
Caveats
- Premium pricing varies by effort setting and channel
- Max-effort configurations consume very high token budgets