Moonshot AILanguage model
Kimi K2 Thinking Turbo
Kimi K2 Thinking Turbo has no overall index value yet because there are too few independent benchmark results.
10out of 100+3 vs. last month
based on two data groups, mostly search
Summary
Kimi K2 Thinking Turbo is listed as a discontinued API model in Kimi’s official catalog. Publisher sources identify no separate Turbo model card or weight release; they point to the distinct Thinking checkpoint with 1 trillion total parameters, 32 billion active parameters, long context, and tool calls.
Moonshot’s catalog marks kimi-k2-thinking-turbo as discontinued from May 25, 2026. The underlying open model card is Kimi K2 Thinking, so the service label and weights remain separate.
All benchmark results
Every published result with its source, test setup and comparison models.
Show all 4 published benchmark results
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
Arena Text Overall human preference rating (Show measurement, test conditions, and source)
- Source value
- 1430
- Score
- 1430
- Metric
- Arena Elo-like human preference rating
- Unit
- elo
- Benchmark version
- Arena Text Overall snapshot 2026-10-15
- Category
- general-human-preference
- Direction
- Higher is better
- Sample
- 63368
- Confidence interval
- 3
- Source type
- official-leaderboard
- Evaluator
- Arena Text leaderboard
- Status
- active
- Retrieved at
- 2026-10-16
- Methodology
- Official benchmark leaderboard. Arena's Oct 15, 2026 Text Overall snapshot reports the exact row label kimi-k2-thinking-turbo at 1430±3, with 63,368 votes. It is rank 116 overall and rank 28 in the open-source filtered view. Moonshot's API announcement independently confirms this literal string as an API model ID. The ranking is human preference across text-to-text tasks.
- Context
- An Elo-like preference rating is not a task-specific accuracy score or a universal capability measure. The displayed rank 28 applies only to the filtered open-source view. Arena links the row to a Kimi K2 Thinking card; the exact serving checkpoint behind the named Turbo API alias is not published.
Arena Text Medicine & Healthcare (Show measurement, test conditions, and source)
- Source value
- 1445
- Score
- 1445
- Metric
- Arena Elo-like human preference rating
- Unit
- elo
- Benchmark version
- Arena Text Medicine & Healthcare snapshot 2026-10-15
- Category
- Medicine & Healthcare
- Direction
- Higher is better
- Sample
- 4027
- Confidence interval
- 10
- Source type
- official-leaderboard
- Evaluator
- Arena Text leaderboard
- Status
- active
- Retrieved at
- 2026-10-18
- Methodology
- Official benchmark leaderboard. Arena row label exactly: Kimi K2 Thinking Turbo, matching the assigned model ID and name. The linked Kimi K2 Thinking model card is not treated as proof of the Turbo backend checkpoint. Serving route/revision and selected adjustment remain unknown.
- Context
- The page exposes Style Control, Factuality, and None as available adjustments but the captured text does not expose which setting is selected. Serving route, exact backend checkpoint, and revision are not reported. Votes are leaderboard comparisons, not independent model trials.
Text Arena Overall (Show measurement, test conditions, and source)
- Source value
- 1430
- Score
- 1430
- Metric
- Score
- Unit
- No unit provided
- Direction
- No direction published
- Source type
- official-leaderboard
- Evaluator
- Arena Text Arena
- Status
- active
- Retrieved at
- 2026-10-18
- Methodology
- Official benchmark leaderboard. Text Arena Overall; no model-generation settings are stated on the leaderboard.
- Context
- The source does not declare a score unit or the meaning of ±. Votes are reported separately and are not treated as evaluated N or as confidence intervals.
Text Arena Creative Writing leaderboard score (Show measurement, test conditions, and source)
- Source value
- 1391
- Score
- 1391
- Metric
- Score
- Unit
- No unit provided
- Category
- Creative Writing
- Direction
- No direction published
- Source type
- official-leaderboard
- Evaluator
- Text Arena Creative Writing leaderboard
- Status
- active
- Retrieved at
- 2026-10-20
- Methodology
- Official benchmark leaderboard. Text Arena Creative Writing view. The exact row label is Kimi K2 Thinking Turbo. The page is dated Oct 15, 2026. The served checkpoint revision is unpublished. The displayed 9,534 votes are leaderboard participation, not independent model samples.
- Context
- The source prints 1391±7 but does not define the score scale, direction, or the meaning of ±. The serving commit is unpublished. The displayed 9,534 votes are leaderboard metadata, not a per-cell sample count.
Popularity compared
How often the model is searched for, used and mentioned, compared with every model of its category.
Popularity index
The 100 most popular models of this category, 0 to 100, higher is more popular
20 of 101 models
As of November 2026 · Source: gradually.ai popularity index
Over time
Weekly values from October 18, 2026 to November 7, 2026
Kimi K2 Thinking TurboSeed 2.0 Lite 13Seed 2.0 Code 13Seed 1.6 11Seed 1.6 Flash 11
Method changes in the timeline, a jump at these points is no market event:
- 1November 1, 2026: Value only with search or usage data, 30-day attention
What is gaining popularity right now, across models, vendors, harnesses and AI subscriptions, is collected in All AI trends.
Similar models
More models of the same kind as Kimi K2 Thinking Turbo, chosen by model line, developer, size, price and release date.
- Kimi K2.5Moonshot AI · same developer
- Kimi K2 ThinkingMoonshot AI · same developer
- Kimi K2 0905Moonshot AI · same developer
- Kimi K2.6Moonshot AI · same developer
- Kimi K2 0711Moonshot AI · same developer
- Kimi K2Moonshot AI · same developer
- Kimi K2.7 CodeMoonshot AI · same developer
- Kimi K2.7 Code HighspeedMoonshot AI · same developer
All specifications
Architecture
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
Frequently asked questions about Kimi K2 Thinking Turbo
What is the knowledge cutoff of Kimi K2 Thinking Turbo?
August 2024, according to Moonshot AI. That is 15 months before release.
Stated by Moonshot AI
How large is the context window of Kimi K2 Thinking Turbo?
262,144 tokens. That is how much text Kimi K2 Thinking Turbo reads at once, about 390 A4 pages.
Page count as a rule of thumb from Artificial Analysis
Can I run Kimi K2 Thinking Turbo locally?
Yes, the weights are open. With 1T parameters it needs a lot of memory, more a server than a laptop.
Stated by Moonshot AI
- Open weights
- yes
- License
- Modified MIT
- Parameters
- 1T
- active per token
- 32B
Sources and data date
Show all 18 sources
Benchmark sources at a glance
Which independent evaluators tested this model. This is coverage, not a quality ranking.
| Evaluator | Test series | Composite indices | Recorded active measurements |
|---|---|---|---|
| LMArena | 0 | 0 | 4 |
Without vendor claims and historical or disputed results; those stay labelled in the single values. 4 measurements are not yet assigned to a test series.
| Type | Evidence and data date |
|---|---|
| Input | Documented modalities. Retrieved October 27, 2026. |
| Output | Documented modalities. Retrieved October 27, 2026. |
| Research status | Research date October 15, 2026. 9 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Unresolved | serving-checkpoint-identity Note The Arena row and Moonshot API announcement share the exact literal model ID kimi-k2-thinking-turbo, so the dated alias-level preference observation is recorded. Arena does not disclose the exact serving checkpoint behind that API alias; no underlying-weight identity is inferred. |
| Unresolved | Turbo backend checkpoint and selected adjustment Original wording The exact Arena row is present, but the linked non-Turbo model card does not identify the Turbo backend checkpoint or the page's selected adjustment. |
| Additional source | Hugging Face (retrieved October 15, 2026) |
| Additional source | Arena Text leaderboard (retrieved October 16, 2026) |
| Additional source | Arena Text leaderboard (retrieved October 18, 2026) |
| Additional source | Arena Text Arena (retrieved October 18, 2026) |
| Additional source | Text Arena Creative Writing leaderboard (retrieved October 20, 2026) |
| Additional source | Model metadata (retrieved October 15, 2026) |
| Additional source | Model metadata (retrieved October 16, 2026)123 |
| Additional source | Kimi API model list · Editorial description reviewed October 2, 2026 |
| Additional source | Kimi K2 Thinking model card · Editorial description reviewed October 2, 2026 |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Ollama: cumulative downloads for the popularity index |
| Additional source | Replicate: cumulative runs for the popularity index |
| Additional source | Google Trending Now: limited feed selection for Germany and the US |
| Additional source | Creative Commons Attribution 4.0 |