Fable 5.1
A large language model that demonstrates strong performance on real-world engineering tasks, exhibiting practical capabilities in large-scale project builds that exceed what its benchmark scores alone would suggest.
Core Facts
The API model identifier for GPT-6 Astra is gpt-6-astra.
Opus 5.5 reaches the level of Fable 5.1 on most tasks, but with approximately 40% lower operating costs and faster speed.
On the DeepSuite benchmark, Grok 4.7 scored 71%, GPT 5.6 Sol scored 72.7%, Fable 5.1 scored 70%, and Astra Max topped the leaderboard at 74.1%.
Timeline (last 90 days)
Jacob said that it was Fable 5 (and later 5.1) that made Spawn's degree of autonomy possible; Opus 5.5 was not sufficient for them.
This ultrathink behavior covers four models: Opus 5.5, Sonnet 5.5, Haiku 5.5, and Fable 5.1.
On Office QA Pro v2 and Spreadsheet Bench 2, Kimi K3.1 Pro both achieved the highest scores, surpassing Claude Opus 5.5, Fable 5.1 and GPT-6 Astra.
Kimi K3.1 Pro costs $2.11 per intelligence task, lower than Fable 5.1's $7.63 and Muse Wonder Max's $4.54.
Using Fable 5.1 on Pro and Team Standard consumes usage credits (an extra charge); on Max and Team Premium it is deducted from the weekly quota, capped at half the quota.
Stephen Robles uses Claude Code (with the Fable 5.1 model) as his preferred AI coding tool rather than Codex.
The participating models included Opus 5.5 and Fable 5.1 accessed via Claude Code, GPT-6-Astra and GPT-6.1-Sol accessed via Codex, and Gemini 4 Argon (High) and Gemini 3.8 Flash (High) accessed via Gemini CLI.
GPT-6-Astra scored 2 points but had 2 strikes; Fable 5.1, Gemini 3.8 Flash and GPT-6.1-Sol each scored 1 point and all accumulated 3 strikes.
P99 latency on Fable 5.1 dropped from 6.6 seconds to 6.4 seconds, a limited long-tail improvement, because 30% of requests still call the model.
Extrapolated at 1 million requests per day over 30 days per month, Fable 5.1 could save about $211,000 per month while Haiku 5.5 saves only about $1,900; this extrapolation is arithmetic based on the measured per-request cost, not another actual measurement, and excludes HotKV's own hosting and operations costs.
40 more timeline events
All Facts (20)
The API model identifier for GPT-6 Astra is gpt-6-astra.
90%VerifiedOpus 5.5 reaches the level of Fable 5.1 on most tasks, but with approximately 40% lower operating costs and faster speed.
80%VerifiedOn the DeepSuite benchmark, Grok 4.7 scored 71%, GPT 5.6 Sol scored 72.7%, Fable 5.1 scored 70%, and Astra Max topped the leaderboard at 74.1%.
65%VerifiedOn Terminal Bench 4.0, Opus 5.5 scored 66.4%, Astra scored 57.9%, Fable 5.1 scored 55.8%, and the previous generation Opus 5 scored 52.3%.
65%Partially VerifiedFable 5.1 processed Venus radar photographs taken by NASA more than thirty years ago into clearer topographic maps.
65%UnverifiedFable 5.1 has stronger coding capabilities.
60%UnverifiedFable 5.1 is a large language model that supports multimodal input and tool calling.
60%UnverifiedFable 5.1 scored 52.6% on the Terminal Bench Science 0.1 benchmark, more than doubling the score of its previous generation.
60%UnverifiedJacob said that it was Fable 5 (and later 5.1) that made Spawn's degree of autonomy possible; Opus 5.5 was not sufficient for them.
50%UnverifiedThis ultrathink behavior covers four models: Opus 5.5, Sonnet 5.5, Haiku 5.5, and Fable 5.1.
50%UnverifiedOn Office QA Pro v2 and Spreadsheet Bench 2, Kimi K3.1 Pro both achieved the highest scores, surpassing Claude Opus 5.5, Fable 5.1 and GPT-6 Astra.
50%UnverifiedKimi K3.1 Pro costs $2.11 per intelligence task, lower than Fable 5.1's $7.63 and Muse Wonder Max's $4.54.
50%UnverifiedUsing Fable 5.1 on Pro and Team Standard consumes usage credits (an extra charge); on Max and Team Premium it is deducted from the weekly quota, capped at half the quota.
50%UnverifiedStephen Robles uses Claude Code (with the Fable 5.1 model) as his preferred AI coding tool rather than Codex.
50%UnverifiedThe participating models included Opus 5.5 and Fable 5.1 accessed via Claude Code, GPT-6-Astra and GPT-6.1-Sol accessed via Codex, and Gemini 4 Argon (High) and Gemini 3.8 Flash (High) accessed via Gemini CLI.
50%UnverifiedGPT-6-Astra scored 2 points but had 2 strikes; Fable 5.1, Gemini 3.8 Flash and GPT-6.1-Sol each scored 1 point and all accumulated 3 strikes.
50%UnverifiedP99 latency on Fable 5.1 dropped from 6.6 seconds to 6.4 seconds, a limited long-tail improvement, because 30% of requests still call the model.
50%UnverifiedThe preserved-thinking check introduced by Fable 5.1 now applies to Opus 5.5 and Sonnet 5.5: thinking blocks are bound to the model and the conversation prefix, and sending thinking blocks back after editing system, tools or earlier turns returns a 400 on accounts created on or after August 31, 2026, while other accounts have them silently dropped.
50%UnverifiedAura offers a variety of frontier models to choose from, including GPT-6 Astra, Fable 5.1, and Opus 5.5, as well as a fully unrestricted automatic mode; users who have already used Claude Code can connect to Aura with one click.
50%UnverifiedIn the advisor pairing mode, each 1 percentage point improvement in performance requires paying about 1.24 times more cost.
50%