← All models
Moonshot AI · modified_mit_license
Kimi K2.6
Kimi K2.6 by Moonshot AI appears in 2 sources with Reasoning at 58.13. Best read for Code quality, Coding, High intelligence.
CreatorMoonshot AI
Release date2026-04-20
Knowledge cutoffNot published
Context262K tokens
Input price$0.95/M tokens
Output price$4/M tokens
Modalitytext + vision
CountryCN
Metrics
All source-backed metrics
LLM Stats Rank12ranking · source_rank
Reasoning58.13reasoning · index_reasoning
Math49.69math · index_math
Coding43.73coding · index_code
Research37.3research · index_search
Vision37.08multimodal · index_vision
Tool calling31.28tool_calling · index_tool_calling
GPQA90.5 %reasoning · gpqa_score
SWE-bench Verified80.2 %coding · swe_bench_verified_score
Code Arena1,547.01 %coding · coding_arena_score
Humanity Last Exam36.4 %reasoning · hle_score
MMMU-Pro80.1 %multimodal · mmmu_pro_score
BrowseComp86.3 %research · browsecomp_score
Toolathlon50 %tool_calling · toolathlon_score
SciCode52.2 %coding · scicode_score
Apex Agents27.9 %agent · apex_agents_score
SWE-bench Pro58.6 %coding · swe_bench_pro_score
CharXiv-R86.7 %multimodal · charxiv_r_score
Context262,144 tokenscontext · context
Speed49.85 c/sperformance · throughput
Latency55,436.36 msperformance · latency
Input price0.95 $/Mpricing · input_price
Output price4 $/Mpricing · output_price
Parameters1,000,000,000,000 paramsmodel · params
Arena Rating1,455.55metric
Arena Rank17metric
Vote Count8,843metric
Evidence
Citations and source overlap
FAQ
How should I read this profile?
Treat this as a source-backed model dossier, not an EvalKit-run verification. The public values are replicated from linked sources and kept source-scoped.
Is Kimi K2.6 verified by EvalKit?
No. EvalKit currently shows 0 verified rows until real run evidence exists.
Why can metrics disagree?
Different sources test different tasks, dates, prompts, and aggregation methods. EvalKit keeps those differences visible instead of merging them into a fake universal score.