--- title: "Grok 4.5 API: facts and benchmarks | Kendr" canonical: "https://kendr.org/models/kc-grok-4.5" date_modified: "2026-08-12" profile_kind: "hosted" --- # Grok 4.5 API model profile Grok 4.5 is a xAI model profile with a 500,000-token context window. The dated catalog snapshot records text, image, file modalities and interfaces for agentic, file input, prompt caching, reasoning, reasoning controls. Grok 4.5 is available under the Kendr alias kc-grok-4.5. Current account availability and customer credit quotes come from Kendr's live public model API and applicable account policy. Grok 4.5 accepts text, image, and file and returns text. The snapshot records agentic, file input, prompt caching, reasoning, reasoning controls, structured output, text generation, tool calling, tools, vision, and web search pricing as capabilities or interfaces. The dated reference snapshot lists $2 input and $6 output per million tokens. Cached input is $0.3 per million tokens. The separately published provider-route reference is $2 / $6 (<200K input; $0.30/M cached). The 2026-08-12 snapshot includes 3 Artificial Analysis indexes and 16 Design Arena categories. Publisher-reported launch results are labeled and are not presented as Kendr measurements; scores from different suites are not treated as interchangeable. Grok 4.5 ranked #30 in the cited trailing-7-day third-party catalog usage dataset as of 2026-08-12. Observed within the cited trailing-7-day third-party catalog usage dataset; this is not global AI market share. Profile data was reviewed for the 2026-08-12 snapshot. Claims retain their source dates, and unavailable fields are shown as unavailable instead of estimated. ## Profile facts - **Profile type:** Kendr API model - **Kendr API alias:** kc-grok-4.5 - **Model developer:** xAI - **Kendr route provider:** xAI - **Context window:** 500K - **Snapshot date:** 2026-08-12 - **Reference catalog ID:** x-ai/grok-4.5 ## Context, modalities, and identity Grok 4.5 accepts text, image, and file and returns text. The snapshot records agentic, file input, prompt caching, reasoning, reasoning controls, structured output, text generation, tool calling, tools, vision, and web search pricing as capabilities or interfaces. - **Model developer:** xAI - **Kendr route provider:** xAI - **Context window:** 500K - **Maximum output:** No verified numeric limit - **Input modalities:** text, image, and file - **Output modalities:** text - **Knowledge cutoff:** Not disclosed ## Capabilities and supported controls Capabilities and parameters are reported from the dated catalog and source records; their presence does not guarantee identical behavior across every provider route. - **Supported parameters:** frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_logprobs, top_p - agentic - file input - prompt caching - reasoning - reasoning controls - structured output - text generation - tool calling - tools - vision - web search pricing ## Dated pricing snapshot The dated reference snapshot lists $2 input and $6 output per million tokens. Cached input is $0.3 per million tokens. The separately published provider-route reference is $2 / $6 (<200K input; $0.30/M cached). Reference figures are dated 2026-08-12; the live Kendr quote can differ by selected provider route, context tier, caching, tools, region, and current rate card. - **Price date:** 2026-08-12 - **Input:** $2 per 1M tokens - **Cached input:** $0.3 per 1M tokens - **Output:** $6 per 1M tokens | Prompt threshold | Input | Cached input | Output | | --- | --- | --- | --- | | From 200,000 prompt tokens | $4 / 1M | $0.6 / 1M | $12 / 1M | ## Benchmark evidence and limitations The 2026-08-12 snapshot includes 3 Artificial Analysis indexes and 16 Design Arena categories. Publisher-reported launch results are labeled and are not presented as Kendr measurements; scores from different suites are not treated as interchangeable. Third-party catalog fields normalized from the cited model dataset; benchmark methodology and coverage differ by source. - **Intelligence Index:** 55.8 - **Coding Index:** 72.4 - **Agentic Index:** 48.9 | Design Arena category | Elo | Win rate | Rank | | --- | --- | --- | --- | | agenticgamedev | 1221 | 56.6% | #5 | | androidnative | 1267 | 67.7% | #3 | | fullstack | 1286 | 65.8% | #3 | | godotgamedev | 1270 | 62.1% | #3 | | htmlslides | 1236 | 56% | #5 | | mobileapps | 1282 | 60.9% | #2 | | python-pptxslides | 1232 | 54.5% | #9 | | webapps | 1253 | 54.3% | #10 | | 3d | 1311 | 50.3% | #11 | | asciiart | 1306 | 58.4% | #6 | | codecategories | 1309 | 53.2% | #9 | | dataviz | 1309 | 54.3% | #11 | | gamedev | 1319 | 53.2% | #10 | | svg | 1270 | 52.2% | #8 | | uicomponent | 1322 | 53.9% | #9 | | website | 1314 | 56.3% | #8 | ## Popularity and market context Grok 4.5 ranked #30 in the cited trailing-7-day third-party catalog usage dataset as of 2026-08-12. Observed within the cited trailing-7-day third-party catalog usage dataset; this is not global AI market share. - **Third-party catalog rank:** #30 - **Global market share:** Not inferred - **Observation date:** 2026-08-12 ## OpenAI-compatible API example ```bash curl https://api.kendr.org/v1/chat/completions \ -H "Authorization: Bearer $KENDR_API_KEY" \ -H "Content-Type: application/json" \ -d '{"model":"kc-grok-4.5","messages":[{"role":"user","content":"Hello"}]}' ``` ## Frequently asked questions ### What is Grok 4.5? Grok 4.5 is a xAI model profile with a 500,000-token context window. The dated catalog snapshot records text, image, file modalities and interfaces for agentic, file input, prompt caching, reasoning, reasoning controls. ### What context window does Grok 4.5 have? The dated profile lists 500K of context. Provider routes, variants, and runtime configuration can impose lower effective limits. ### What benchmark evidence is available for Grok 4.5? The 2026-08-12 snapshot includes 3 Artificial Analysis indexes and 16 Design Arena categories. Publisher-reported launch results are labeled and are not presented as Kendr measurements; scores from different suites are not treated as interchangeable. Third-party catalog fields normalized from the cited model dataset; benchmark methodology and coverage differ by source. ### What is the Kendr API alias for Grok 4.5? Use kc-grok-4.5 as the model value. The live public model API and signed-in catalog remain authoritative for route availability and customer credit quotes. ### How is Grok 4.5 priced on Kendr? The dated reference snapshot lists $2 input and $6 output per million tokens. Cached input is $0.3 per million tokens. The separately published provider-route reference is $2 / $6 (<200K input; $0.30/M cached). Kendr applies one 5% markup to configured provider model cost; the live quote and settled routing receipt are authoritative for a request. ## Sources and evidence dates - [Grok 4.5 third-party catalog record](https://openrouter.ai/x-ai/grok-4.5) — catalog, checked 2026-08-12 - [Third-party model catalog methodology](https://openrouter.ai/docs/guides/overview/models) — methodology, checked 2026-08-12 - [xAI official model documentation](https://docs.x.ai/developers/pricing) — primary, checked 2026-08-12 - [Artificial Analysis capability indices methodology](https://artificialanalysis.ai/methodology/capability-indices) — benchmark, checked 2026-08-12 - [Design Arena leaderboard and methodology](https://www.designarena.ai/leaderboard) — benchmark, checked 2026-08-12 - [Grok 4.5 provider or route reference](https://docs.x.ai/developers/pricing) — primary, checked 2026-08-12 Live Kendr operational availability is published separately at https://api.kendr.org/api/public/models.