← comparisons · head to head
GPT-6 Astra vs Claude Fable 5.1
Of the 6 public leaderboards that rate both, GPT-6 Astra scores higher on 2 and Claude Fable 5.1 on 4. On the ArtificialWatch Index, which takes the median of each model's percentile across the boards that score it, GPT-6 Astra is at 89.2 and Claude Fable 5.1 at 95.3 (out of 100).
Scores as published by each board; table composed September 24, 2026.
Leaderboard by leaderboard
| Board | GPT-6 Astra | Claude Fable 5.1 | Higher |
|---|---|---|---|
| LMArena Elo (community votes) | 1,792.2 | 1,754.7 | GPT-6 Astra |
| LiveBench global average (0-100) | 82.2 | 83.4 | Claude Fable 5.1 |
| SimpleBench AVG@5 % (private set) | 86.5 | 86.6 | Claude Fable 5.1 |
| LLM Stats LLM Stats Score · composite (0-100), their rendered top 15 | 59.5 | 54.5 | GPT-6 Astra |
| Vals AI Vals Index accuracy % · weighted finance + coding tasks | 66.6 | 68.8 | Claude Fable 5.1 |
| Artificial Analysis intelligence index · measured runs only | 52.7 | 53.4 | Claude Fable 5.1 |
Raw scores are each board's own scale, so compare within a row, not across rows. Full table: the bench index.
GPT-6 Astra at a glance
- Developer
- OpenAI
- Status
- Live since September 3, 2026
Live — Sep 3, Trusted Access enterprises first; our own sweep read the plain id on OpenRouter at the base sticker Sep 4 21:16 UTC (openai/gpt-6-astra, $10/$50), and the Fast tier the same night (vercel/openai/gpt-6-astra-fast, $20/$100) · $10/$50 per 1M, $1 cached, 1.05M ctx / 128K out · AA Coding Agent Index 67 (≈ Opus 5, Fable 5), Intelligence Index 61 (= GPT-5.6 Sol, 5 below Fable 5.1).
GPT-6 Astra is out — Fable 5.1's sticker, trusted access first, public API within days
2026-09-03 · OpenAI shipped GPT-6 Astra today. Its own words, from the model documentation: "Our most capable model, built for the hardest end-to-end work" — complex reasoning, coding, computer use, research and document creation.
OpenAI is treating Astra as its first Critical cyber model — and slowing it down, 24 hours after the leaks said next week
2026-08-07 · Twenty-four hours ago both accounts we track said Astra was launching next week. Today OpenAI published that it is slowing the model down, and the reason outranks any launch calendar.
Claude Fable 5.1 at a glance
- Developer
- Anthropic
- Status
- Live since September 1, 2026
Live - sticker price held at $10/$50, but cache reads cut 75% to $0.25 = ~25% cheaper typical, up to ~45% agentic · 1M ctx · effort low-max.
Claude Fable 5.1 is live — the sticker price never moved, and the cache read fell 75%
2026-09-01 · Claude Fable 5.1 shipped today, alongside Claude Mythos 5.1. Our radar detected it at 17:52:22 UTC and confirmed it on the next sweep 60 seconds later.
Fable 5.1's August window has closed — the market now prices September at 78%, and one claim in the leak can't be checked at all
2026-08-20 · A claim going round today says Fable 5.1 is in pre-release testing, targeted at September, and running inside Anthropic's own products under the Fable 5 name. One part of that is checkable, one part is checkable and slipping, and one part is — by construction — invisible to anything we can measure.
GPT-6 Astra vs Claude Fable 5.1: quick answers
Which is better, GPT-6 Astra or Claude Fable 5.1?
It depends on what you measure. Of the 6 public leaderboards that rate both, GPT-6 Astra scores higher on 2 and Claude Fable 5.1 on 4. GPT-6 Astra's widest lead is on LLM Stats; Claude Fable 5.1's widest lead is on LiveBench. On the ArtificialWatch Index, which takes the median of each model's percentile across the boards that score it, GPT-6 Astra is at 89.2 and Claude Fable 5.1 at 95.3 (out of 100).
Which is newer, GPT-6 Astra or Claude Fable 5.1?
GPT-6 Astra. It went live on September 3, 2026, 2 days after Claude Fable 5.1 (September 1, 2026).
Who makes GPT-6 Astra and Claude Fable 5.1?
GPT-6 Astra is from OpenAI; Claude Fable 5.1 is from Anthropic.
Where do these numbers come from?
Each score is the leaderboard's own published figure, read by ArtificialWatch (table composed September 24, 2026). Boards measure different things, so raw scores are only comparable within a row.
More comparisons
Watch free: Chrome push the moment a new model answers on a public API, an email about 15 minutes later. Also: the bench index · the wire