Pinned "latest" alias for the ChatGPT-style chat-tuned variant of GPT-5.3. Tracks current revision; pricing matches GPT-5.3 base.
GPT-5.3 Chat (latest) is TheRouter's pinned alias for OpenAI's current GPT-5.3 chat-tuned model revision. It is meant for teams that want the latest GPT-5.3 chat behavior without rotating application code every time OpenAI refreshes the underlying checkpoint. TheRouter exposes it through the standard OpenAI-compatible chat completions endpoint as openai/gpt-5.3-chat-latest.
Operationally, this is a high-context text-and-vision chat model: 400K context, 128K max completion, text and image input, text output, and support for temperature, top_p, max_tokens, tools, tool_choice, response_format, and stop. Pricing is inherited from the active TheRouter catalog at $1.89 per million input tokens and $15.12 per million output tokens. Treat it as a production chat target when you value up-to-date GPT-5.3 behavior more than perfectly frozen repeatability.
The important caveat is in the name: latest aliases can move. For regression-sensitive workflows, evaluation baselines, regulated approvals, or deterministic release gates, pin a fixed sibling such as openai/gpt-5.3-codex, openai/gpt-5.4, or another explicit model ID after testing. For general assistant, multimodal support, and tool-calling chat where current quality matters, this alias is the convenient route.
OpenAI operates gpt-5.3-chat-latest as a moving chat-tuned alias for the latest GPT-5.3 chat revision.
TheRouter serves the same alias through the OpenAI-compatible /v1/chat/completions endpoint; TheRouter does not freeze the upstream revision behind the latest ID.
| Type | Rate |
|---|---|
| Input | $1.89 / 1M tokens |
| Output | $15.12 / 1M tokens |
| TheRouter model ID | openai/gpt-5.3-chat-latestdevelopers.openai.com ↗ | verified |
| Model role | Pinned latest alias for the chat-tuned GPT-5.3 revision | verified |
| Context window | 400,000 tokens | verified |
| Max completion | 128,000 tokens | verified |
| Input / output modalities | Text + image input → text output | verified |
| Supported parameters | temperature, max_tokens, top_p, tools, tool_choice, response_format, stopplatform.openai.com ↗ | verified |
| TheRouter pricing | $1.89/M input tokens, $15.12/M output tokens | verified |
| Training cutoff | Not publicly disclosed | unknown |
| License | Proprietary — OpenAI API terms and policiesopenai.com ↗ | verified |
| Benchmark | Distribution | Score | Source |
|---|---|---|---|
Public reasoning benchmark No official reasoning benchmark table specific to the gpt-5.3-chat-latest alias is published in the sources reviewed for this page. | — | Not publicly disclosed | — |
Public coding benchmark Do not copy GPT-5.3-Codex coding benchmark claims onto this chat alias without a source that names the alias. | — | Not publicly disclosed | — |
Public multimodal benchmark The catalog confirms image input, but reviewed public sources do not publish a multimodal score for this exact latest alias. | — | Not publicly disclosed | — |
Use the global api.therouter.ai endpoint shown below for new integrations; the legacy China accelerated endpoint is retired.
curl https://api.therouter.ai/v1/chat/completions -H "Content-Type: application/json" -H "Authorization: Bearer $THE_ROUTER_API_KEY" -d '{
"model": "openai/gpt-5.3-chat-latest",
"messages": [
{"role": "user", "content": "Summarize the key points from this input."}
]
}'Call GPT-5.3 Chat Latest through TheRouter's OpenAI-compatible chat completions endpoint.
curl https://api.therouter.ai/v1/chat/completions \
-H "Authorization: Bearer $THEROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"openai/gpt-5.3-chat-latest","messages":[{"role":"user","content":"Summarize the risks in this rollout plan."}]}'TheRouter added a dedicated page for openai/gpt-5.3-chat-latest so operators can see its alias semantics, 400K context, vision input, tool-calling support, and the reproducibility caveat of using a latest model ID.
It is TheRouter's model ID for the current chat-tuned GPT-5.3 latest alias. Use it when you want up-to-date GPT-5.3 chat behavior through an OpenAI-compatible API and accept that the upstream revision can move over time.
Not as the only baseline. Latest aliases are convenient for current quality but can change behavior. For regression tests, approvals, and published benchmark numbers, pin a fixed explicit model after running your own evals.
Yes. TheRouter lists the model as text+image input to text output. Send image_url content blocks through the normal chat completions message format.
GPT-5.3 Chat Latest is the chat-tuned latest alias for broad assistant workloads. GPT-5.3-Codex is the coding-specialist model for long-running software engineering and terminal-heavy agent workflows. If the task is primarily code editing, start with Codex; if it is broad chat with vision and tools, this alias is the simpler choice.
No. The examples are structurally standard OpenAI-compatible requests, but no paid snippet verification run was performed for this scheduled curation pass. Treat the snippets as integration templates until an operator funds verification.
| source | URL | retrieved | |
|---|---|---|---|
| TheRouter model ID | developers.openai.com ↗ | 2026-08-11 | verified |
| Model role | — | — | verified |
| Context window | — | — | verified |
| Max completion | — | — | verified |
| Input / output modalities | — | — | verified |
| Supported parameters | platform.openai.com ↗ | 2026-08-11 | verified |
| TheRouter pricing | — | — | verified |
| Training cutoff | — | — | unknown |
| License | openai.com ↗ | 2026-08-11 | verified |
| Public reasoning benchmark | developers.openai.com ↗ | 2026-08-11 | unknown |
| Public coding benchmark | developers.openai.com ↗ | 2026-08-11 | unknown |
| Public multimodal benchmark | platform.openai.com ↗ | 2026-08-11 | unknown |
| TheRouter documents GPT-5.3 Chat Latest as a dedicated model page | developers.openai.com ↗ | 2026-08-11 | verified |
| What is openai/gpt-5.3-chat-latest? | developers.openai.com ↗ | 2026-08-11 | to verify |
| Does GPT-5.3 Chat Latest support images? | platform.openai.com ↗ | 2026-08-11 | to verify |
| How is it different from GPT-5.3-Codex? | openai.com ↗ | 2026-08-11 | to verify |