Smaller, more cost-effective version of GPT-5.1-Codex.
GPT-5.1-Codex-Mini is the smaller, faster, and more cost-effective member of OpenAI's GPT-5.1 Codex family — built specifically for quick code edits, interactive coding sessions, and cost-sensitive agentic workflows. Released on November 13, 2025 alongside GPT-5.1 and GPT-5.1-Codex, it shares the same 400K context window and 128K max output as its siblings but at roughly one-fifth the price: $0.25/M input tokens and $2/M output tokens.
Like all Codex-family models, GPT-5.1-Codex-Mini is accessed through Codex surfaces (CLI, IDE extension, cloud sandbox, code review) or via the OpenAI API. It supports the same parameter set as GPT-5.1-Codex — temperature, max_tokens, tools, tool_choice, response_format, reasoning, stop — including adaptive reasoning effort with 'none' through 'high' levels. From a TheRouter operator's perspective, the model is listed as active and reachable through the standard /v1/chat/completions endpoint at the attractive $0.25/$2 price point.
Use GPT-5.1-Codex-Mini when you need fast, cost-effective coding assistance for interactive editing, quick PR reviews, or lightweight agentic tasks that don't require the frontier reasoning of GPT-5.1-Codex or GPT-5.1-Codex-Max. For long-running autonomous coding sessions (24+ hours), front-end GPT-5.1-Codex-Max is the better choice; for general reasoning and chat, route to openai/gpt-5.1.
| Type | Rate |
|---|---|
| Input | $0.270 / 1M tokens |
| Output | $2.16 / 1M tokens |
| Release date | 2025-11-13openai.com ↗ | verified |
| OpenAI API model id | gpt-5.1-codex-miniopenai.com ↗ | verified |
| Model type | Cost-effective agentic coding model — smaller, faster variant of GPT-5.1-Codex optimized for interactive editing and lightweight coding automationopenai.com ↗ | verified |
| Reasoning effort levels | Supports 'none', 'low', 'medium', 'high' reasoning effort via reasoning_effort parameter; defaults to 'none' for lowest latencyopenai.com ↗ | verified |
| Input modalities | Text, Image (vision)openai.com ↗ | verified |
| License | Proprietary (OpenAI API terms of service / ChatGPT subscription) | verified |
| Benchmark | Distribution | Score | Source |
|---|---|---|---|
SWE-bench Verified OpenAI has not published SWE-bench scores for GPT-5.1-Codex-Mini. The model is positioned as a cost-effective coding companion, not a benchmark-chasing frontier model. For SWE-bench performance, use GPT-5.1-Codex-Max (77.9%) or GPT-5.2-Codex. | — | 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-5.1-codex-mini",
"messages": [
{"role": "user", "content": "Summarize the key points from this input."}
]
}'GPT-5.1-Codex-Mini uses the standard chat completions endpoint. For best latency in interactive coding tasks, set reasoning_effort to 'none' (default) or 'low'. For more complex coding tasks, 'medium' provides a good balance of accuracy and speed.
curl https://api.therouter.ai/v1/chat/completions \
-H "Authorization: Bearer $THEROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"openai/gpt-5.1-codex-mini","messages":[{"role":"user","content":"Write a function to merge two sorted arrays in TypeScript"}],"reasoning_effort":"low"}'OpenAI announced GPT-5.1 in the API platform, alongside the Codex-family models gpt-5.1-codex and gpt-5.1-codex-mini. GPT-5.1 introduces adaptive reasoning (including 'none' mode), extended prompt caching for 24 hours, and new apply_patch and shell tools for coding. GPT-5.1-Codex-Mini is positioned as the cost-effective coding companion at $0.25/M in tokens.
GitHub rolled out GPT-5.1, GPT-5.1-Codex, and GPT-5.1-Codex-Mini in Copilot public preview for Pro, Pro+, Business, and Enterprise users. Codex-Mini was made available in VS Code, JetBrains, Xcode, Eclipse, and Copilot CLI across all modes (chat, ask, edit, agent).
OpenAI retired GPT-5.1 Instant, GPT-5.1 Thinking, and GPT-5.1 Pro from ChatGPT on March 11, 2026. The gpt-5.1-codex-mini API model id was not affected — the model remains accessible via the OpenAI API and through TheRouter at api.therouter.ai.
GPT-5.1-Codex-Mini is a smaller, faster, and significantly more cost-effective version of GPT-5.1-Codex. It is priced at $0.25/M input and $2/M output tokens — roughly one-fifth the cost of GPT-5.1-Codex ($1.25/$10). It shares the same 400K context window, 128K max output, and full parameter set (tools, reasoning_effort, JSON mode). The trade-off is reduced benchmark performance: Mini is designed for quick, interactive coding tasks rather than frontier-level autonomous agentic coding. For 24+ hour coding sessions and compaction, use GPT-5.1-Codex-Max.
No. Compaction (native multi-window context management) is a feature exclusive to GPT-5.1-Codex-Max. Codex-Mini operates within a single 400K context window. For long-running autonomous sessions that require compaction, route to openai/gpt-5.1-codex-max via TheRouter.
Technically yes, but OpenAI recommends using Codex-family models exclusively for coding and agentic development workflows. For general reasoning, chat, or non-coding tasks, route to openai/gpt-5.1 (general-purpose GPT-5.1) or openai/gpt-5-mini (budget option) via TheRouter for better results.
For fastest interactive coding in the IDE, use reasoning_effort='none' (default). For simple code generation, fix, or review tasks, use 'low'. For debugging, code review, or multi-step refactoring, use 'medium'. The model does not support 'xhigh' — that is exclusive to GPT-5.1-Codex-Max.
| source | URL | retrieved | |
|---|---|---|---|
| Release date | openai.com ↗ | 2026-05-28 | verified |
| OpenAI API model id | openai.com ↗ | 2026-05-28 | verified |
| Model type | openai.com ↗ | 2026-05-28 | verified |
| Reasoning effort levels | openai.com ↗ | 2026-05-28 | verified |
| Input modalities | openai.com ↗ | 2026-05-28 | verified |
| License | — | — | verified |
| SWE-bench Verified | openai.com ↗ | 2026-05-28 | unknown |
| OpenAI releases GPT-5.1, GPT-5.1-Codex, and GPT-5.1-Codex-Mini for developers | openai.com ↗ | 2026-05-28 | verified |
| GPT-5.1-Codex-Mini available in GitHub Copilot public preview | github.blog ↗ | 2026-05-28 | verified |
| GPT-5.1 models retired from ChatGPT (API unaffected) | help.openai.com ↗ | 2026-05-28 | verified |
| How is GPT-5.1-Codex-Mini different from GPT-5.1-Codex? | openai.com ↗ | 2026-05-28 | to verify |
| Does GPT-5.1-Codex-Mini support compaction or multi-window context? | openai.com ↗ | 2026-05-28 | to verify |
| Can I use GPT-5.1-Codex-Mini for non-coding tasks? | openai.com ↗ | 2026-05-28 | to verify |
| What reasoning effort should I use for GPT-5.1-Codex-Mini? | openai.com ↗ | 2026-05-28 | to verify |