Claude Fable 5.2
Top of the board at 65, and at $8/$40 a fifth cheaper than 5.1 — the strongest model you can call today
Benchmark scores
How it is measured →Ranked 1 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 65.0 | 100% | |
| AA Intelligence Index v4.3.2 | — | — | not indexed |
| GPQA Diamond | — | — | not indexed |
| Terminal-Bench 4.0 Terminal-agent tasks | 59.6 | 89% | −7.6 pts |
| LiveBench Design | — | — | not indexed |
| Vals Index | — | — | not indexed |
Assessment
First on AA Intelligence Index v5 at 65, five points clear of second-placed Gemini 4 Carbon, and the only model at the top that anyone can call. Terminal-Bench 4.0 (AA's own run) is 59.6%, level with Opus 5.5 and still behind Kimi K3.1 Pro and Carbon. 20% cheaper than 5.1 ($10/$50 → $8/$40), though at full effort it costs $26.54 per task, about 2.6 times Opus 5.5; even its high setting scores 62, higher than most models manage at max. Ships with cyber and bio safeguards that hand blocked requests to Opus, and Mythos 5.2 on the same base model stays restricted to vetted organisations and is not listed here. No published results for GPQA or LiveBench, so both are left blank.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| GPT-6 AstraOpenAI | 56.9 | $20.00 | 81 | -8.1 |
| Claude Opus 5.5Anthropic | 59.1 | $8.00 | 94 | -5.9 |
| Gemini 4 CarbonGoogle | 60.0 | $8.00 | 88 | -5.0 |
| Muse WonderMeta | 44.0 | $8.00 | 143 | -21.0 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboardRelated benchmarks
Data sources & methodology
Full methodology & sources →Every number on this page comes from published evaluations. We run no tests of our own and accept no vendor-submitted scores. Main sources:
Evaluation data in the wider industry also comes from public datasets, leaderboard platforms and their evaluation harnesses, private in-house sets, production logs and user feedback, and synthetic data. This site's current rankings use only the parts that are public, independently checkable and clearly defined.