Kimi K3 vs GPT-5.6 Sol
Side-by-side comparison of API pricing, specs, benchmarks, and capabilities
| Specification |
by Moonshot
|
GPT-5.6 SolIntro price
by OpenAI
|
|---|---|---|
| Overview | ||
| Status | Current flagship Open weights | Current flagship |
| Released | Jul 16, 2026 | Jul 9, 2026 |
| Pricing per million tokens | ||
| Input | $3.00/Mtok | $4.00/Mtok |
| Output | $15.00/Mtok | $20.00/Mtok |
| Blended avg | $6.00/Mtok | $8.00/Mtok |
| Cached input | $0.300/Mtok | $0.400/Mtok |
| Price basis | — | OpenAI labels this promotional pricing and states it is "available at least through November 21, 2026". That is a floor, not a reversion date — OpenAI publishes no date on which the rate goes back up, so no reversion is claimed here and no countdown is implied. Until 2026-08-21 this model billed $5 input / $0.50 cached input / $30 output per 1M tokens. OpenAI's rate card also prices this model by context length: the tracked figure is the short-context tier, and prompts over the short-context threshold bill at the long-context tier of $8 input / $0.80 cached input / $10 cache write / $30 output per 1M tokens. Cache writes bill at 1.25x the uncached input rate. |
| Specifications | ||
| Context window | 1M tokens | 1M tokens |
| Parameters | Proprietary | Proprietary |
| Speed (TPS) | — | — |
| Modalities | ||
| Input |
textimage
|
textimage
|
| Benchmarks source: Vellum LLM Leaderboard | ||
| Avg benchmark score | 83.3 | 83.8 |
| Perf / dollar | 13.9 | 10.5 |
| GPQA Diamond | 93.5 | 94.6 |
| SWE-Bench Verified | 93.4 | 96.2 |
| Terminal-Bench 2.1 | 88.3 | 88.8 |
| BrowseComp | 91.2 | 92.2 |
| Humanity's Last Exam | 56 | 47.2 |
| HumanEval | 76.8 | 96.2 |
| MCP Atlas | 84.2 | — |
| OSWorld-Verified | — | 62.6 |
| ARC-AGI 2 | — | 92.5 |
| AutoBench | — | 18.1 |
| Independent composite scores each on its own scale — not part of the average above | ||
| AA Intelligence Index | 59.7 | 60.9 |
| AA Agentic Index | 54.3 | 57.8 |
| AA-Omniscience Index (−100–100) | 19.7 | 22 |
| GDPval-AA v2 | 1668 ELO | 1710.1 ELO |
| LMSYS Chatbot Arena | 1485.3 ELO | 1482.7 ELO |
| Providers | ||
| Available from |
Moonshot — $3.00/$15.00/Mtok
|
OpenAI — $4.00/$20.00/Mtok
|
Cost at scale
1M tokens · 50/50 input/output| Volume | Kimi K3 | GPT-5.6 Sol | Savings |
|---|---|---|---|
| 1M tokens | $6 | $8 | $2 (25%) |
| 10M tokens | $60 | $80 | $20 (25%) |
| 100M tokens | $600 | $800 | $200 (25%) |
| 1000M tokens | $6000 | $8000 | $2000 (25%) |
When to pick which
distilled from the pricing and spec data above- cost dominates: $6.00/Mtok blended vs $8.00 — 25% less on the same 50/50 token mix
- you want open weights — self-host it, fine-tune it, or exit the API entirely; GPT-5.6 Sol is closed
- raw capability matters more than price: 83.8 vs 83.3 average benchmark score
Price movement
recorded changes since we began tracking each model · last checked Sep 2, 2026| Model | Changes | Latest move | Old → new $/M |
|---|---|---|---|
| Kimi K3 | 0 | — no move | Stable since we began tracking — no change recorded |
| GPT-5.6 Sol | 1 | price cut on Aug 21, 2026 | $5/$30 → $4/$20/Mtok |
The cheaper option here — Kimi K3 costs $6.00/Mtok blended on Moonshot.
Summary
Kimi K3 by Moonshot costs $3.00/Mtok input and $15.00/Mtok output, with a 1M-token context window. It supports text, image input.
GPT-5.6 Sol by OpenAI costs $4.00/Mtok input and $20.00/Mtok output, with a 1M-token context window. It supports text, image input.
On a blended cost basis, Kimi K3 is 25% cheaper than GPT-5.6 Sol.
On benchmarks, GPT-5.6 Sol scores higher (83.8 vs 83.3) on average. In terms of value, Kimi K3 has better performance per dollar (13.9 vs 10.5).
Note: Pricing is per million tokens. Actual costs vary with usage patterns, prompt caching, and batch discounts. Always verify against official provider pricing pages.