Kimi K3.1 Pro
- Family
- Kimi
- Released
- 2026-10-23
- Context
- 1M
- Parameters
- 2800B total; 104B active
- Architecture
- Mixture of Experts
- Specialization
- general
- Openness
- Proprietary
- License
- ProprietaryRestricted
- Weights
- Coming soon
- Code
- Unknown
Cheapest of 2 routes · Moonshot AI Kimi · cache read $0.200
About
Kimi K3.1 Pro is the first model in Moonshot AI's K3.1 series, a post-trained version of Kimi K3. It has a 1,048,576-token context window, always-on reasoning with low, high and max effort, image and video input, tool calling and strict JSON Schema structured output. Released 2026-10-23 on Kimi.ai, Kimi Work, Kimi Code and the Kimi API; Moonshot says full weights will follow by 2026-11-04.
Providers(2)
Compare all 2Compare API pricing across 2 tracked provider routes for input and output tokens, batch, and cached reads when available.
| Provider | Input / 1M | Output / 1M | Cache | Route |
|---|---|---|---|---|
| Moonshot AI Kimi | $2.00 | $10.00 | read $0.200 / 5m $2.00 / 1h $4.00 | Serverless |
| Vercel AI Gateway | $2.00 | $10.00 | read $0.200 | Serverless |
Capabilities
Benchmark peer barsfor RAG
No task-mapped benchmark peers are available for this model yet.
Benchmark scores(16)
| Benchmark | Score | Version | Evaluation | Source |
|---|---|---|---|---|
| 65.7 | Terminal-Bench 4.0; Claude-Code harness; max reasoning effortObserved 2026-10-23 | — | ||
| 75.5 | DeepSWE v1.1; Claude-Code harness; max reasoning effortObserved 2026-10-23 | — | ||
| 77.1 | Kimi Code Bench 2.0 (internal); Claude-Code harness; max reasoning effortObserved 2026-10-23 | — | ||
| 70.0 | HLE-Full with tools; max reasoning effortObserved 2026-10-23 | — | ||
| 51.1 | HLE-Full without tools; max reasoning effortObserved 2026-10-23 | — | ||
| 94.2 | diamond; max reasoning effortObserved 2026-10-23 | — | ||
| 92.1 | BrowseComp; context compaction at 200K tokens; max 500 steps; max reasoning effortObserved 2026-10-23 | — | ||
| 43.2 | Agents' Last Exam v1.1 (ALL, 122 tasks); max reasoning effortObserved 2026-10-23 | — | ||
| 79.6 | OSWorld V2.1; Kimi Agent harness; max reasoning effortObserved 2026-10-23 | — | ||
| 89.0 | CharXiv RQ; no tools; max reasoning effortObserved 2026-10-23 | — | ||
| 84.1 | MMMU-Pro; official protocol; no tools; max reasoning effortObserved 2026-10-23 | — | ||
| 34.6 | GDP.pdf vision, no tools; strict pass rate; five-run mean; max reasoning effortObserved 2026-10-23 | — | ||
| 20.7 | ExploitGym; KimiCode harness; 6-hour budget; max reasoning effortObserved 2026-10-23 | — | ||
| 95.4 | BabyVision w/ python; four-run mean; max reasoning effortObserved 2026-10-23 | — | ||
| 90.2 | ChartoGraphy w/ python; five-run mean; max reasoning effortObserved 2026-10-23 | — | ||
| 63.0 | ZeroBench-main w/ python Pass@5; max reasoning effortObserved 2026-10-23 | — |
Migration checks
No linked migration route is available for this model yet.
Cheapest of 2 routes · Moonshot AI Kimi · cache read $0.200