Kimi K2.6
Moonshot AI · Launched Apr 2026
Moonshot's open agentic model; SWE-bench Verified 80.2%, SWE-bench Pro 58.6%, Terminal-Bench 2.0 66.7%. Sustains 4,000+ tool calls over 13-hour sessions.
Benchmark performance
OpenAI-verified 500-issue subset of SWE-Bench. Approaching saturation in 2026 - most frontier models clear 80%+.
Held-out, contamination-resistant CLI tasks driven end-to-end in a real terminal. Version 2.1 is the 2026 standard for terminal autonomy.
Harder, contamination-resistant successor to SWE-Bench Verified: real GitHub issues with held-out tests. Where coding headroom remains.
Other coding-focused agents
The 9 agents in this category, ranked by peak benchmark.
| Agent | Maker | Launch | Peak | Pricing |
|---|---|---|---|---|
| Kimi K2.5OSS | Moonshot AI | 2026-01 | 1410.0 | Open weights |
| Claude Code | Anthropic | 2025-02 | 88.6 | Claude Max / API |
| Codex CLI | OpenAI | 2025-04 | 83.4 | ChatGPT / API |
| SWE-AgentOSS | Princeton + Stanford | 2024-04 | 74.0 | Open source (MIT) |
| Gemini CLIOSS | 2025-06 | 70.7 | Free tier + API | |
| GLM-5.1OSS | Z.ai | 2026-04 | 58.4 | Open weights |
| OpenCodeOSS | OpenCode | 2025-06 | — | Open source |
| AiderOSS | Aider | 2023-05 | — | Open source (Apache-2) |
Recent coverage
2026-06-01
NVIDIA Nemotron 3 Ultra: 550B Open-Weight Model Challenges GLM, Kimi
2026-05-23
Cerebras Hits 981 Tokens/sec on 1T-Parameter Kimi K2.6, Claims 6.7× GPU Cloud Speedup
2026-05-11
CoreWeave Tops Kimi K2.6 Inference Speed
2026-04-24
DeepSeek V4-Pro: 1.6T parameters, open weights, undercuts rivals 10x
Quick facts
- Type
- Coding-focused
- Maker
- Moonshot AI
- Launch
- 2026-04-01
- Open source
- Yes
- Pricing
- Open weights
- Benchmarks scored
- 4
- Article mentions
- 4
- Rank in category
- #4 of 9