Benchmark Grok 4.5 before moving coding agents off Claude or GPT
xAI has put Grok 4.5 on the API for coding, agentic tasks, and knowledge work. The useful move now is not trust the launch claims, but run your own repo evals against its $2/$6 pricing and 500K-context availability.
Confirmed: xAI lists Grok 4.5 as available from July 8 for coding, agentic tasks, and knowledge work. The headline numbers are $2 per 1M input tokens and $6 per 1M output tokens, with configurable reasoning effort. Treat performance claims as a starting point, not a migration plan, until your own coding-agent evals pass.
What changed
xAI published Grok 4.5 and the dedicated model documentation now points developers to the new model. OpenRouter also lists x-ai/grok-4.5 with a 500K context window, the same $2/$6 headline pricing, and a July 8 release date.
| Model | Best fit | Access | Cost/status | Caveat |
|---|---|---|---|---|
| Grok 4.5 | Coding agents, STEM, knowledge work | xAI API, OpenRouter | $2 input / $6 output per 1M tokens | Validate long-context and tool-use behavior yourself |
| Grok Build 0.1 | xAI coding CLI workflows | Early access | Separate model | Not the same launch |
Why this is early
TestingCatalog surfaced Grok 4.5 traces before the official launch window, including web references that pointed to a near-term release. That made it useful as a watch signal, but the publishable confirmation is now the xAI developer release note plus public model availability on OpenRouter.
Key takeaways
- Grok 4.5 is a confirmed API model, not just a social leak.
- xAI positions it around coding, agentic work, and knowledge tasks.
- OpenRouter availability makes quick side-by-side tests easier for teams already routing models.
- Pricing is aggressive enough to test on real repo tasks, but benchmark claims still need independent proof.
Availability and access
Developers can try Grok 4.5 through the xAI API and OpenRouter. xAI says reasoning effort can be set to low, medium, or high, with high as the default. Before committing volume, check the xAI model page for context-tier pricing, rate limits, data controls, and region/account restrictions.
Why it matters
Grok 4.5 is worth testing if your current agent stack spends heavily on coding, code review, migration, or long-context repository work. The practical LinkLoot move is to run the same failing-test repair, multi-file refactor, and documentation-update tasks across Grok 4.5, Claude, GPT, and your current fallback model. See LinkLoot's /guides/ai-agent-tools hub for agent-stack selection context.
What to verify before you act
- Confirm
grok-4.5is enabled in your xAI account or router. - Compare actual cached and uncached cost on a real workflow, not a single prompt.
- Test tool calling, patch quality, and regression handling on your own repo.
- Check enterprise data controls before sending proprietary code.
- Watch for a system card, benchmark detail, and changelog updates.
Source check
Confirmed by: xAI's Grok 4.5 announcement, xAI model documentation, and OpenRouter model availability.
Early signal / context: TestingCatalog spotted pre-launch traces; Axios gives independent launch context and market framing. LinkLoot will treat official benchmark detail, safety documentation, or wider product availability as update triggers.
Yes. xAI lists Grok 4.5 in its July 8 developer release notes, and OpenRouter lists the model as available.