Grok 4.5
Grok 4.5 is worth evaluating for coding, rag, and agents when its provider route and context window match the workload.
Use it for
- Teams evaluating coding, rag, and agents
- Workloads that can use a 500k context window
- Buyers comparing 3 tracked provider routes
Do not use it for
- Workloads where another current model has stronger sourced task evidence
- Family
- Grok 4
- Released
- 2026-07-08
- Context
- 500k
- Specialization
- code
- Openness
- Proprietary
- License
- ProprietaryCommercial use: conditional
- Weights
- Not released
- Code
- Unknown
- Training
- Pretrained
Cheapest of 3 routes · OpenRouter · cache read $0.500
About
Grok 4.5 is xAI/SpaceXAI's July 2026 flagship model for coding, agentic tasks, and knowledge work. It is publicly available as grok-4.5 through the xAI API and Grok Build, supports configurable low/medium/high reasoning, function calling, web search, X search, code execution, prompt caching, and a 500k-token context window. xAI tiered pricing applies at a 200K prompt threshold: <=200K prompts at $2.00/$0.50 cached/$6.00 per 1M tokens (input/cached input/output), and prompts above 200K at $4.00/$1.00/$12.00; EU API-console availability was not yet live on launch day.
Grok 4.5 is a proprietary model in the Grok 4 family. The structured metadata tracks a 500k-token context window, multimodal input, reasoning, function calling, tool use, and code execution. This page tracks provider routes through xAI Console, OpenRouter, and Vercel AI Gateway, with the cheapest tracked route listed at $2 input and $6 output per 1M tokens. Headline tracked benchmarks include DeepSWE 1.0 62.0, DeepSWE 1.1 53.0, and Terminal-Bench 2.1 83.3.
Top use-case fit: coding, agents, and build tasks
Coding
Q/$ D1 relevant benchmark in the decision map.
RAG
Included by capability and metadata signals in the decision map.
Agents
Included by capability and metadata signals in the decision map.
Provider price ladder
Compare all 3Compare API pricing across 3 providers for input and output tokens, batch, and cached reads when available.
| Provider | Input / 1M | Output / 1M | Cache | Route |
|---|---|---|---|---|
| OpenRouter | $2.00 | $6.00 | read $0.500 | Serverless |
| Vercel AI Gateway | $2.00 | $6.00 | read $0.500 | Serverless |
| xAI Console | $2.00 | $6.00 | read $0.500 | Serverless |
Available via routers & gateways(1)
Capabilities
Benchmark peer barsfor Coding
Benchmark scores(7)
| Benchmark | Score | Version | Evaluation | Source |
|---|---|---|---|---|
| DeepSWE 1.0 | 62.0 | DeepSWE 1.0, Pass@1, Datacurve-created eval with each model provider's harnesses by AA (xAI first-party claim)Observed 2026-07-08 | — | Source |
| DeepSWE 1.1 | 53.0 | DeepSWE 1.1, Pass@1, mini-swe-agent harness run by Datacurve (xAI first-party claim)Observed 2026-07-08 | — | Source |
| Terminal-Bench 2.1 | 83.3 | Terminal Bench 2.1 (xAI first-party reported chart; harness not specified)Observed 2026-07-08 | — | Source |
| SWE-bench Pro | 64.7 | SWE Bench Pro resolve rate (xAI first-party reported chart)Observed 2026-07-08 | — | Source |
| CursorBench | 66.7 | CursorBench 3.2Observed 2026-07-18 | Configuration: Grok 4.5 High* Harness: CursorBench 3.2 productized Cursor-agent workflow Evaluator: Cursor Confidence: confirmed Cost/task: $1.51 Tokens/task: 19,521 Steps/task: 33 Qualification: Cursor discloses a training-data advantage; do not use this result for a neutral ranking claim. Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful. | Source |
| CursorBench | 65.4 | CursorBench 3.2Observed 2026-07-18 | Configuration: Grok 4.5 Medium* Harness: CursorBench 3.2 productized Cursor-agent workflow Evaluator: Cursor Confidence: confirmed Cost/task: $1.54 Tokens/task: 18,914 Steps/task: 34 Qualification: Cursor discloses a training-data advantage; do not use this result for a neutral ranking claim. Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful. | Source |
| CursorBench | 63.5 | CursorBench 3.2Observed 2026-07-18 | Configuration: Grok 4.5 Low* Harness: CursorBench 3.2 productized Cursor-agent workflow Evaluator: Cursor Confidence: confirmed Cost/task: $1.22 Tokens/task: 15,841 Steps/task: 31 Qualification: Cursor discloses a training-data advantage; do not use this result for a neutral ranking claim. Notes: Cursor vendor-reported result; not independently reproducible. Results are subject to variance, and small score differences may not be statistically meaningful. | Source |
Migration checks
No linked migration route is available for this model yet.
API versions
grok-4.5Frequently asked questions
What is the context window of Grok 4.5?
Grok 4.5 has a context window of 500k tokens.
How much does Grok 4.5 cost?
Grok 4.5 is available at $2.00/1M input tokens through xAI Console.
When was Grok 4.5 released?
Grok 4.5 was released on 2026-07-08.
Which providers offer Grok 4.5?
Grok 4.5 is available from 3 providers: xAI Console, OpenRouter, Vercel AI Gateway.
What benchmarks has Grok 4.5 been tested on?
Grok 4.5 has been evaluated on 7 benchmarks, including DeepSWE 1.0, DeepSWE 1.1, Terminal-Bench 2.1, SWE-bench Pro, CursorBench.
Cheapest of 3 routes · OpenRouter · cache read $0.500