Kimi K2.7 Code HighSpeed vs GPT-5.3 Codex
Side-by-side comparison of API pricing, specs, benchmarks, and capabilities
| Specification |
by Moonshot
|
by OpenAI
|
|---|---|---|
| Overview | ||
| Status | Current coding Open weights | Current coding |
| Released | Jun 12, 2026 | Feb 24, 2026 |
| Pricing per million tokens | ||
| Input | $1.90/Mtok | $1.75/Mtok |
| Output | $8.00/Mtok | $14.00/Mtok |
| Blended avg | $4.95/Mtok | $7.88/Mtok |
| Cached input | $0.380/Mtok | $0.175/Mtok |
| Specifications | ||
| Context window | 262K tokens | 400K tokens |
| Parameters | Proprietary | Proprietary |
| Speed (TPS) | — | — |
| Modalities | ||
| Input |
textimagevideo
|
textimage
|
| Benchmarks sources: Moonshot model card / Vellum LLM Leaderboard | ||
| SWE-Bench Verified | 50 vendor | 85 vendor |
| Avg benchmark score | — | 71 |
| Perf / dollar | — | 9 |
| Terminal-Bench 2.1 | — | 77.3 |
| OSWorld-Verified | — | 64.7 |
| HumanEval | 89 vendor | — |
| Independent composite scores each on its own scale — not part of the average above | ||
| AA Intelligence Index | — | 44.3 |
| LMSYS Chatbot Arena | — | 1416 ELO |
| Providers | ||
| Available from |
Moonshot — $1.90/$8.00/Mtok
|
OpenAI — $1.75/$14.00/Mtok
|
Cost at scale
1M tokens · 50/50 input/output| Volume | Kimi K2.7 Code HighSpeed | GPT-5.3 Codex | Savings |
|---|---|---|---|
| 1M tokens | $4.95 | $7.88 | $2.93 (37.2%) |
| 10M tokens | $49.5 | $78.75 | $29.25 (37.1%) |
| 100M tokens | $495 | $787.5 | $292.5 (37.1%) |
| 1000M tokens | $4950 | $7875 | $2925 (37.1%) |
The cheaper option here — Kimi K2.7 Code HighSpeed costs $4.95/Mtok blended on Moonshot.
Summary
Kimi K2.7 Code HighSpeed by Moonshot costs $1.90/Mtok input and $8.00/Mtok output, with a 262K-token context window. It supports text, image, video input.
GPT-5.3 Codex by OpenAI costs $1.75/Mtok input and $14.00/Mtok output, with a 400K-token context window. It supports text, image input.
On a blended cost basis, Kimi K2.7 Code HighSpeed is 37.1% cheaper than GPT-5.3 Codex.
The two aren't directly comparable on average benchmark score: GPT-5.3 Codex has published per-benchmark results, while Kimi K2.7 Code HighSpeed does not yet.
Note: Pricing is per million tokens. Actual costs vary with usage patterns, prompt caching, and batch discounts. Always verify against official provider pricing pages.