Kimi K3 vs GPT-5.3 Codex
moonshotai/kimi-k3 · openai/gpt-5.3-codex
As of November 2, 2026, Kimi K3 is the cheaper API on UberLLM: $0.0017 per million input tokens and $0.0781 per million output, versus $0.0555 / $0.4438 for GPT-5.3 Codex, about 83% less when output is 3 tokens for every input token. These are the best prices on the market right now. Both work with one OpenAI-compatible key and UberLLM adds no fee. Choose Kimi K3 to cut cost; choose GPT-5.3 Codex when its capabilities fit your task better.
Kimi K3 vs GPT-5.3 Codex, side by side
| Spec | Kimi K3 | GPT-5.3 Codex |
|---|---|---|
| Price · input / 1M | $0.0017 | $0.0555 |
| Price · output / 1M | $0.0781 | $0.4438 |
| Context length | 1,000,000 | 400,000 |
| Max output tokens | 131,072 | 128,000 |
| Input modalities | text, image | text, image |
| Output modalities | text | text |
| Tools / functions | Yes | Yes |
| Vision (image input) | Yes | Yes |
| Reasoning | Yes | Yes |
| % off list | 99.3% off | 96.8% off |
| First token, median (7 days) | 3.3 s (1,158 requests) | — |
Best prices on UberLLM right now. No UberLLM fee on requests; adding credit by card costs 5.5% + $0.50. Updated November 2, 2026. A price more than 70% below list is what buyers were billed, not an error. Why so cheap?
Kimi K3 vs GPT-5.3 Codex — FAQ
Is Kimi K3 or GPT-5.3 Codex cheaper?
Kimi K3 is cheaper on a blended token mix, by about 83%. A request costs what its provider charged, UberLLM adds no fee, and every request is listed with its cost in Activity.
What's the price difference between Kimi K3 and GPT-5.3 Codex?
Kimi K3: $0.0017 in / $0.0781 out per 1M tokens. GPT-5.3 Codex: $0.0555 in / $0.4438 out per 1M tokens. That's roughly a 83% blended difference. Prices are live; UberLLM adds no fee.
Can I use both Kimi K3 and GPT-5.3 Codex with one API?
Yes. Both are served through a single OpenAI-compatible endpoint — set the base URL to https://api.uberllm.dev/v1 and your UberLLM key, then switch between them by changing the model id. Anthropic SDKs work via https://api.uberllm.dev/v1/messages.