Key facts
- SpaceXAI launched Grok 4.5 on July 8.
- Grok 4.5 is priced at $2 per million input tokens and $6 per million output tokens.
- Elon Musk claimed the model is comparable to Anthropic's Opus 4.7 but faster.
- Benchmark results show Grok 4.5 scoring 53% on DeepSWE 1.1 and 64.7% on SWE Bench Pro.
- Grok 4.5 demonstrated efficiency, using significantly fewer output tokens per task than Opus 4.8.
- European access to Grok 4.5 is anticipated in mid-July.
SpaceXAI has launched its latest AI model, Grok 4.5, on July 8, positioning it as a cost-effective alternative to leading models from Anthropic and OpenAI. The model is priced at $2 per million input tokens and $6 per million output tokens, significantly less than comparable offerings.
Elon Musk announced the release on X, stating that Grok 4.5 is roughly comparable to Anthropic's previous flagship, Opus 4.7, but offers much greater speed. While benchmarks show mixed results, with Grok 4.5 scoring 53% on DeepSWE 1.1 and 64.7% on SWE Bench Pro, the company highlights its efficiency. For instance, Grok 4.5 used an average of 15,954 output tokens to complete tasks on SWE Bench Pro, compared to 67,020 tokens for Anthropic's Opus 4.8.
The model was trained using tens of thousands of Nvidia GB300 GPUs on SpaceXAI's Colossus supercomputer. It is available via API and on Hermes and Grok build, offering half a million tokens of context. European users are expected to gain access in mid-July.