SPCX Stock Records 3-Day Slide: Elon Musk Admits Anthropic’s Fable Is 'Definitely Better' Than Grok 4.5

Musk said Grok 4.5 could improve weekly as SpaceXAI works to close the gap with Anthropic and OpenAI in an increasingly crowded AI race.

• Musk said Fable is “definitely better” than Grok 4.5, but added that most users do not need Fable-level capability for everyday tasks.
• Musk also said the next Grok release could bring a “step-change improvement” as SpaceXAI applies lessons learned from Tesla, SpaceX, Neuralink, and the Boring Company's engineering problems.
• Grok 4.5 is priced below several flagship rivals, at $2 per million input tokens and $6 per million output tokens.

SpaceX (SPCX) shares extended their losing streak for a third straight session on Wednesday as investors weighed SpaceXAI’s Grok 4.5 launch, with CEO Elon Musk also acknowledging that Anthropic’s Fable remains ahead of the newly released model in terms of capabilities.

SPCX stock fell nearly 1% on Wednesday, with shares also down 8% so far this week. 

SpaceXAI Challenges Anthropic, OpenAI On Price

The remark from Musk came just as SpaceXAI rolled out Grok 4.5, its latest AI model and the company’s first major public release since going public. The launch was pitched as a major step for developers, engineers, coders, researchers, and other knowledge workers, but Musk’s own comments have framed the model as a practical, cost-efficient challenger rather than the outright leader in the AI race.

“In fairness, Fable is definitely better than Grok 4.5, but most tasks don’t require Fable-level capability,” Musk said on X. SpaceXAI said Grok 4.5 is a model “built for real-world engineering,” saying it excels in large codebases and can handle long-running tasks across multiple repositories, hundreds of skills, and a variety of tools.

Musk added that the next release could mark another major leap, saying: “Next month’s release will be another step-change improvement, as we close the loop on solving real-world engineering problems at Tesla, SpaceX, Neuralink and Boring Company.”

Musk earlier called Grok 4.5 an “Opus-class model,” saying it was comparable to Anthropic’s Opus 4.7 but faster, more token-efficient, and lower-cost. 

SpaceXAI said Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens, making it cheaper than several rival high-end models if its performance holds up. Anthropic’s Opus 4.7 costs $5 per million input tokens and $25 per million output tokens, while OpenAI’s top-tier Sol model is priced at $5 per million input tokens and $30 per million output tokens. OpenAI’s lower-cost Luna model is closer, at $1 for input and $6 for output. 

Benchmarks Put Grok 4.5 Behind Fable

SpaceXAI’s launch benchmarks showed that Grok 4.5 performed strongly in some software engineering tests but trailed Anthropic’s top models in others. On DeepSWE 1.1, which measures how reliably AI systems close real software bugs, Grok 4.5 scored 53%, behind Claude Opus 4.8, GPT-5.5, and Fable 5. On SWE Bench Pro, Grok 4.5 performed better, scoring 64.7%, ahead of GPT-5.5 but still behind Opus 4.8 and Fable 5.

Musk also said that Grok 4.5’s current performance does not fully reflect SpaceXAI’s hardware potential. “Grok 4.5 is not yet using our internally developed C/C++ inference software that exactly maps to the GB300 hardware,” Musk said on X. “Doubling or more of the current speed is probably achievable.” The company trained Grok 4.5 using tens of thousands of Nvidia GB300 GPUs inside its Colossus supercomputer in Memphis, which reportedly has capacity across more than 200,000 GPUs.

Musk Says Grok 4.5 Will Improve Weekly

Musk further said that the pace of progress is accelerating, suggesting Grok 4.5 will improve rapidly after launch. “Rate of improvement is accelerating,” Musk said. “Users should notice a meaningful improvement in the usefulness of the Grok Build harness with our V9 foundation model (aka Grok 4.5) every week.” The weekly improvement claim may be important for SpaceX investors watching whether SpaceXAI can close the gap with Anthropic and OpenAI.

The Grok 4.5 launch comes during a crowded week for AI releases, with OpenAI preparing the public rollout of GPT-5.6 and Anthropic pushing ahead with its own powerful model lineup. Both companies are also moving toward public listings, making the race for enterprise AI dominance more intense.

How Do Retail Traders Feel About SPCX?

On Stocktwits, retail sentiment for SPCX was ‘bearish’ amid ‘extremely low’ message volume.

One user said, “$SPCX I think this could get to $110-120 on a market pullback.”

View this Stocktwits post

Another user said, “$SPCX still too much hype!  No data of the business until the first earnings report. So it’s skeptical & speculation for now, imo. The company is valuation at $135 per share.”

View this Stocktwits post

For updates and corrections, email newsroom[at]stocktwits[dot]com.<