Live in Kimi Code · Sep 11, 2026 Model ID unchanged — kimi-for-coding

K2.8Kimi PREVIEW

The most efficient thinker in the Kimi family. Coding and agent capabilities closing in on flagship K3 — with adjustable thinking effort and a 1M-token context window on every membership tier, rolled out with zero configuration.

thinking · low / high / max context · 1M tokens access · all tiers migration · none needed
≈K3
Coding performance
1M
Context, all tiers
max
Default thinking
Scroll
01 — Capability matrix

Everything K2.8 Preview
brings to the table

Announced in the official Kimi Code changelog on September 11, 2026. Every claim below traces to that release note — no speculation, no rumor.

CAP / 01

Near-flagship coding

Officially described as "performance close to K3" — Moonshot's flagship. Coding and agent capabilities are improved across the board over K2.7 Code, with gains concentrated in long-horizon, multi-step engineering work.

CAP / 02

Radically efficient thinking

Significantly more efficient thinking than K2.7 Code — itself a model that cut reasoning-token usage 30% versus K2.6. Less overthinking, more doing: fewer tokens burned per solved task.

CAP / 03

Adjustable thinking effort

The same three thinking levels as K3 — low / high / max — with max as the default. Turn thinking off entirely and requests are still served, now by K2.8 in instant mode.

CAP / 04

1M context, every tier

A context window of up to one million tokens — available across all membership tiers. The flagship K3 reserves its 1M window for higher tiers; K2.8 Preview democratizes it.

CAP / 05

Zero-configuration rollout

The model ID is unchanged — still kimi-for-coding. Every client, IDE plugin, CI pipeline, and third-party tool keeps working exactly as before, now on a stronger model.

CAP / 06

The instant-mode engine

With thinking switched off, requests to the K3 series and K2.8 Preview are both served by K2.8 Preview (no thinking). One model now powers both fast answers and deep deliberation across the lineup.

02 — Thinking effort

Dial the deliberation
up or down

K2.8 Preview introduces K3-grade thinking control to the mainstream lineup. Pick a level to see how the trade-off shifts.

Thinking effort · default = max
DEFAULT · FULL DELIBERATION

Maximum reasoning depth

The out-of-the-box setting. K2.8 reasons at full depth before acting — the mode behind its near-K3 results on long-horizon coding and agentic tasks.

Reasoning depth
100%
Latency
higher
Reasoning tokens
100%
With thinking switched off, K2.8 Preview serves requests directly with no reasoning trace — it also becomes the engine for no-thinking K3 requests.
03 — Context window

One million tokens,
no tier gating

Whole monorepos, marathon sessions, and massive document sets — in a single uninterrupted conversation. On K3, 1M requires a higher membership tier. On K2.8 Preview, it's everywhere.

0 tokens
K2.8 Preview
1,000,000
K3 · lower tiers
256K*
~750K
words of natural language ≈
~1,500
typical source files ≈
~20
novels of context in one session ≈
sessions killed mid-task (well, fewer)

* K3 unlocks 1M only on higher tiers (Allegretto and above); K2.8 Preview ships 1M on every tier. ≈ figures are order-of-magnitude approximations for scale, not official measurements.

04 — Lineup comparison

Where K2.8 Preview sits

Three generations of Kimi coding models, side by side — from the K2.7 Code specialization to the 2.8-trillion-parameter K3 flagship.

DimensionK2.7 CodeK2.8 PreviewK3 (flagship)
RELEASEJune 2026 · GASep 11, 2026 · rolling out2026 · open-sourced
POSITIONINGSpecialist coding modelNear-K3 coding for everyoneFlagship · 2.8T params
THINKINGThinking on / offlow / high / max (default max)low / high / max
CONTEXTLong-context1M — all tiers1M · higher tiers only
ARCHITECTURENot disclosed for previewKDA hybrid linear attention + Attention Residuals
VISION INPUTNative visual understanding
ACCESSKimi for CodingAll Kimi Code users, zero configModerato membership+
05 — Lineage

The road to K2.8

April 2026

Kimi K2.6 Open weights

The open-weight generation: topped SWE-Bench Pro at 58.6% in its launch comparison and introduced the Thinking / Instant split that would define the family.

June 2026

Kimi K2.7 Code

+10.4% Program-Bench, +11.4% MCP Mark Verified, +76.2% SWE Marathon over K2.6 — while reasoning-token usage dropped 30%. The efficiency story starts here.

September 11, 2026

K2.8 Preview You are here

Fully rolled out in Kimi Code. Near-K3 coding performance, K3-grade thinking controls, 1M context on every tier — under the unchanged model ID kimi-for-coding.

Flagship · 2026

Kimi K3

The 2.8-trillion-parameter ceiling: KDA hybrid linear attention, native visual understanding, and benchmark wins over Opus 4.8 and GPT-5.5. K2.8 Preview closes most of the gap at every tier.

06 — Known unknowns

What Moonshot hasn't
said yet

The preview label is honest. As of September 15, 2026, these details remain undisclosed by official sources — treat anything claiming them with suspicion.

Moonshot has not published K2.8 Preview's parameter count or architectural details. The only official architecture disclosure in the family is K3's: 2.8 trillion parameters on KDA hybrid linear attention with Attention Residuals. Any site claiming exact K2.8 specs is extrapolating.

The changelog describes K2.8's gains qualitatively — "performance close to K3," "improved across the board," "significantly more efficient thinking" — without SWE-Bench, Program-Bench, or MCP Mark figures. Hard numbers are expected with the stable release or a technical report.

Because the model ID didn't change, existing kimi-for-coding billing carries over — but Moonshot hasn't published K2.8-specific pricing or quota announcements. Current plan limits apply as before.

Preview implies refinement in flight: behavior may shift as Moonshot tunes the model toward a stable K2.8. The changelog notes no breaking changes — same ID, same interfaces — but pinning expectations to the changelog rather than anecdotal benchmarks is the safe move.

The official changelog covers Kimi Code. Secondary reports (Pandaily, Sept 14, 2026) say the same model also reached Kimi Work users — marked as third-party coverage here, not confirmed by the changelog we sourced.

07 — Sources

Every claim, traceable