Kimi K2.6 vs Kimi K3
| Benchmark | Kimi K2.6 | Kimi K3 |
|---|---|---|
| Humanity's Last Exam | 37.5 | 46.9 |
| Terminal-Bench 4.0 | 0.5 | 12.6 |
| AutomationBench-AA | 13 | 58.3 |
| MMMU-Pro (vision) | 79.4 | 80.5 |
Specs
| Spec | Kimi K2.6 | Kimi K3 |
|---|---|---|
| Input, per 1M | $0.95 | $3 |
| Output, per 1M | $4 | $15 |
| Cost per task? | $0.71 | $2.00 |
| Speed, tok/s | 98 | 55 |
| Context | 256K | 1M |
| Weights | Open | Open |
| Released | Apr 2026 | Jul 2026 |
| Memory (4-bit) | 616 GB | 1,668 GB |
| Licence | modified-mit | kimi-k3 |
| Commercial use | Restricted | Restricted |
| See images | ✓ | ✓ |
| Read PDFs | Partial | Partial |
| Hear audio | — | — |
| Watch video | — | ✓ |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | ✓ |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | Partial |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Kimi K2.6 → Kimi K3
Memory (4-bit): 616 GB → 1,668 GB
Sampling parameters are fixed›
temperature, top_p, n and penalty values are fixed; omit them from requests.
No public image URLs›
Vision input rejects public image URLs; use base64 or ms:// file references in a content array.
Web search is being updated›
Web search is not recommended for near-term production workflows; avoid depending on it in production.
Thinking mode is always enabled›
K3 always has thinking enabled; control effort with the top-level reasoning_effort field using low, high or max.
Kimi K3 → Kimi K2.6