SpaceXAI Unveils Grok 4.5 as Opus-Class Model
모델 출시 | Thu Jul 09 2026 00:00:00 GMT+0000 (Coordinated Universal Time) | 1 sources
SpaceXAI released Grok 4.5, its first model since going public, positioning it as a rival to Opus 4.7 with emphasis on token efficiency and low cost.
Analysis
[SpaceXAI] unveiled Grok 4.5 [1]
- First model release since going public
- Supports general knowledge work including coding
- office tasks
- research
- and writing
- Claims 2x token efficiency compared to existing leading models
[Elon Musk] positioned Grok 4.5 as an Opus-class model [1]
- Self-evaluated as roughly on par with Anthropic's Opus 4.7
- Presented faster speed and lower cost as competitive advantages
- Cited positive feedback from beta test customers
[Grok 4.5 pricing] priced at input $2 / output $6 per million tokens [1]
- Opus 4.7 costs input $5 / output $25 per million tokens
- OpenAI Sol costs input $5 / output $30 per million tokens
- OpenAI Luna costs input $1 / output $6 per million tokens
[Benchmark performance] released benchmark metrics at levels similar to competing models [1]
- Performance figures close to top-tier models from competitors
- However
- results fall slightly short of best-in-class
- Real-world token efficiency expected to serve as the practical differentiator
[AI industry dynamics] intensified competition with OpenAI's GPT 5.6 release scheduled the same week [1]
- OpenAI to unveil GPT 5.6 on Thursday
- OpenAI describes it as its 'strongest model yet'
- Release had been delayed due to security concerns from the Trump administration