Grok 4.5 vs Fable 5: Opus-level coding at a fraction of the price
Reviewers are calling Grok 4.5 a bigger deal than Fable 5: Opus-level coding at a fraction of the price.
Transcript
Reviewers are calling Grok 4.5 a bigger deal than Fable 5: Opus-level coding at a fraction of the price.
Grok 4.5 arrived in early July from xAI and SpaceX. It runs on a fresh 1.5 trillion parameter base trained with data from Cursor, the coding tool they acquired.
Here is the surprise. Grok was never known for coding. Reviewers found 4.5 now lands in the same class as Opus 4.8 and GPT 5.5, a huge jump from 4.3.
It also shines on the visual side. In hands-on tests, reviewers built animated websites and one-shot apps with it, and preferred its landing page design over GPT 5.6 Sol.
But the headline is price. In one reviewer's thousand dollar test, the same app cost 189 dollars on Grok, versus 730 on Fable 5 and 327 on GPT 5.6 Sol.
The verdict from reviewers: on benchmarks it sits just below Fable 5 and GPT 5.6, but the price makes it their new daily workhorse. Suddenly xAI is a serious contender.
Quick note: we make Inboxsmith, an AI receptionist that never misses a business call. Get more at inboxsmith dot com.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- A reviewer who calls himself a Grok convert says Grok 4.5 is a huge deal and can act like an AI co-founder.(Grok 4.5 is a bigger deal than Fable 5)
- Grok 4.5 has Opus-level intelligence at a fraction of the cost and time, described as roughly a tenth of the cost of Fable.(Grok 4.5 is a bigger deal than Fable 5)
- In a head to head landing-page build, the reviewer preferred the Grok 4.5 page over the GPT 5.6 Soul page on design and copy.(Grok 4.5 is a bigger deal than Fable 5)
- Grok was never a frontier model for coding, but Grok 4.5 is surprisingly good and in line with Opus 4.8 and GPT 5.5 on coding.(Grok 4.5 is surprisingly good...)
- On artificial analysis, Grok jumped from the bottom of the benchmarks at 4.3 to the front, in line with Opus 4.8 and GPT 5.5, at 4.5.(Grok 4.5 is surprisingly good...)
- Grok 4.5 API pricing is about 2 dollars per million input tokens and 6 dollars per million output tokens, cheaper than Opus 4.8 and GPT 5.5.(Grok 4.5 is surprisingly good...)
- Grok 4.5 completed a task in a Rust and TypeScript codebase that Fable 5 failed six or seven times in a row, and did it faster.(Grok 4.5 is surprisingly good...)
- Grok 4.5 was released by SpaceX AI together with Cursor, and there is now a serious third player: Anthropic, OpenAI, and SpaceX AI.(Grok 4.5 is surprisingly good...)
- Grok 4.5 is surprisingly good at building animated websites inside Cursor, even for non-developers.(Grok 4.5 is SURPRISINGLY Good at Animated Websites)
- In a 1000-dollar side-by-side test off one spec, the same app cost about 189 dollars on Grok 4.5, 730 dollars on Fable 5, and 327 dollars on GPT 5.6 Soul.(Elon Musk's Grok 4.5 Killed Fable 5 and GPT-5.6 Sol in Testing)
- The reviewer judged Grok 4.5 the fastest and the best application in the test, and named it his reigning champion and first recommendation.(Elon Musk's Grok 4.5 Killed Fable 5 and GPT-5.6 Sol in Testing)
- Grok 4.5 is based on a newly pre-trained V9 foundation model of 1.5 trillion parameters, and SpaceX AI acquired Cursor's parent Anysphere in June with Cursor data feeding training.(Grok 4.5 explained in 8min..)
- On the artificial analysis index Grok 4.5 pulls ahead of open models but lands just shy of state-of-the-art models like Fable 5 and GPT 5.6.(Grok 4.5 explained in 8min..)
- Grok 4.5 API pricing of 2 dollars in and 6 dollars out per million tokens undercuts frontier US labs, comparable to lower-tier models like Sonnet 5.(Grok 4.5 explained in 8min..)
- Fable 5 was released June 9, GPT 5.6 was announced June 26, and Grok 4.5 followed, all within about 30 days.(Grok 4.5 explained in 8min..)