Model Follower.

Up to date

Rebuilt
2026-10-25 06:54 UTC
Release radar
117 on watch
Hub snapshot
2026-10-25
Re-checked
daily · every number sourced and dated

Models › Kimi-K2.7 family

Kimi-K2.7-Code

Hugging Face ↗ Follow changes

moonshotai Kimi (Modified MIT) (modified-mit) out 4 months ago

1.03T parameters (8 of 384 experts per token), images and text in, text out, 256K tokens of context. 13 providers serve it from $0.67 per million input tokens.

Verified 2026-10-25 Rank #109 in LLMs · score 98.1 method v1 Sources 4 Downloads 30d 90.6k · #324 of 619 tracked

Can I run it?616GB at INT4Estimateneeds a multi-GPU nodemethod
What does it cost?$0.67/1M incheapest of 12Inceptronint42026-10-24
Can I use it?Kimi (Modified MIT) (modified-mit)MIT plus a display-the-model-name attribution requirement for products above 100M MAU or US$20M monthly revenue; the exact wording is in each repo's LICENSElicence file2026-10-24
Do people use it?90.6kdownloads, 30 days#324 of 619 we track · below the median+3 likes, downloads −18% this week1,424 likesHugging Face2026-10-24

Worth reading

Picked by rule from 12 tracked pieces: the lab's own post, the most shared, a field report, the latest.

The lab's own post

Kimi K2.7 Code

Kimi K2.7 Code is an open-source programming model from Moonshot AI featuring agent capabilities and a 1T parameter MoE architecture with 32B activated parameters per token.

kimi.com 2026-06-13 4 HN pts · 1 comments 539 words
Most discussed · 417 HN pts

Kimi K2.7 Code is generally available in GitHub Copilot

Kimi K2.7 Code is generally available in GitHub Copilot as the first open-weight model offered as a selectable option in the Copilot model picker.

github.blog 2026-07-01 417 HN pts · 185 comments 317 words
Latest · 4 months ago

Kimi K2.7 Code

Kimi K2.7 Code is a dedicated coding model that significantly improves instruction compliance and long-horizon coding performance compared to K2.6.

  • The Vision model has no limit on the number of images, but the request body size must not exceed 100M.
  • For very large videos, the file upload method is mandatory because URL-formatted images are not supported.
platform.kimi.ai 2026-06-13 4 HN pts · 1 comments 753 words

12 best of 12 tracked

The 5 picks above, plus these 7. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

  1. DiscussionKimi-K2.7-Codetwitter.com · 2026-06-13 · 5 HN pts
  2. ArticleKimi K2.7 Code Review: Open Weights Enter Copilot | Awesome Agentsawesomeagents.ai · 2026-07-13
  3. ArticleKimi K2.7-Code: Open-Weight Agentic Coding Model Examined - Beagleheybeagle.com · 2026-07-11
  4. ArticleKimi K2.7-Code Review: Open-Weights 1T MoE Coding Model Deep Diveopen-techstack.com · 2026-06-20
  5. ArticleKimi K2.7 Code: Moonshot's Coding-First Fork of a 1T MoE Base, Now with 30% Less Thinking — Redeepredeep.blog · 2026-06-12
  6. ArticleKimi K2.7-Code cuts thinking tokens 30%venturebeat.com · undated
  7. ArticleKimi K2.7-Code is Out. The Open-Source Coding Model That Thinks Lessdishantsharma.hashnode.dev · undated

Where to run it

Live hosted endpoints, cheapest first, USD per million tokens. Ties break on output price, then provider name.

12 providers from $0.67

12 providers with a live price, from $0.67 to $0.95 per million input tokens; the median is $0.89. One dot per provider, cheapest in teal.

$0.010$0.10$1.00$10.00Inceptron: $0.67 in / $3.35 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeDeepInfra: $0.68 in / $3.40 out per 1M · observed 2026-10-25 · https://huggingface.co/api/models/moonshotai/Kimi-K2.7-Code?expand[]=inferenceProviderMappingCoreWeave: $0.71 in / $3.50 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeStreamLake: $0.71 in / $3.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeVenice: $0.75 in / $3.50 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeSiliconFlow: $0.86 in / $3.80 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeNovita: $0.91 in / $3.84 out per 1M · observed 2026-10-25 · https://huggingface.co/api/models/moonshotai/Kimi-K2.7-Code?expand[]=inferenceProviderMappingNebius: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeGMICloud: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeAlibaba: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeCloudflare: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeMoonshot AI: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codecheapest $0.67median $0.89
$0.010$0.10$1.00$10.00Inceptron: $0.67 in / $3.35 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeDeepInfra: $0.68 in / $3.40 out per 1M · observed 2026-10-25 · https://huggingface.co/api/models/moonshotai/Kimi-K2.7-Code?expand[]=inferenceProviderMappingCoreWeave: $0.71 in / $3.50 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeStreamLake: $0.71 in / $3.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeVenice: $0.75 in / $3.50 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeSiliconFlow: $0.86 in / $3.80 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeNovita: $0.91 in / $3.84 out per 1M · observed 2026-10-25 · https://huggingface.co/api/models/moonshotai/Kimi-K2.7-Code?expand[]=inferenceProviderMappingNebius: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeGMICloud: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeAlibaba: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-codeCloudflare: $0.95 in / $4.00 out per 1M · observed 2026-10-24 · https://openrouter.ai/moonshotai/kimi-k2.7-code+1cheapest $0.67 · median $0.89 (white tick)
#Cheapest 5In /1MOut /1MContextQuantSeen
1 Inceptron $0.67 $3.35 256K int4 Oct 24
2 DeepInfra $0.68 $3.40 256K — Oct 25
3 CoreWeave $0.71 $3.50 256K int4 Oct 24
4 StreamLake $0.71 $3.00 250K unknown Oct 24
5 Venice $0.75 $3.50 250K int4 Oct 24

All 13 providers →

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

Widely served (13 API providers); 28 quantized builds for local use.

  • As a service13Alibaba, Cloudflare, CoreWeave, DeepInfra +9
  • On own hardware28quantized builds
  • Built on1311 fine-tunes · 0 adapters · 2 merges
  • In apps49Hugging Face Spaces

Top Spaces running it: smolagents/ml-intern ♥ 423 · akhaliq/Kimi-K2.7-Code ♥ 9 · avagridworkit/moonshotai-Kimi-K2.7-Code ♥ 1

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Needs an 8×80 GB node even at INT4

About 616 GB at INT4, 1.23 TB at FP8, 2.46 TB at BF16: weights only, so leave room for the KV cache. Every expert must be resident, so total parameters count.

16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 616 GB of weights (estimate; KV cache and activations extra)INT4616 GBFP8: about 1.23 TB of weights (estimate; KV cache and activations extra)FP81.23 TBBF16: about 2.46 TB of weights (estimate; KV cache and activations extra)BF162.46 TB
16 GB card64 GB Mac2×80 GB8×80 GBINT4: about 616 GB of weights (estimate; KV cache and activations extra)INT4616 GBFP8: about 1.23 TB of weights (estimate; KV cache and activations extra)FP81.23 TBBF16: about 2.46 TB of weights (estimate; KV cache and activations extra)BF162.46 TB

Estimate params × bytes per param × 1.2 · how we estimate

The story so far

Weights, every dated article, and repo commits on one rail.

Most coverage came in the first two weeks

Teal is the weights (2026-06-11), amber triangles are 9 dated articles (each links to it), grey ticks are repo commits.

JulAugSepOct2026-06-15: 1 repo commit2026-06-12: 1 repo commit2026-06-11: 2 repo commitsKimi K2.7 Code is generally available in GitHub Copilot · 2026-07-01Kimi K2.7 Code · 2026-06-13Kimi K2.7 Code · 2026-06-13Kimi-K2.7-Code · 2026-06-13Kimi K2.7 Code Review: Open Weights Enter Copilot | Awesome Agents · 2026-07-13Kimi K2.7-Code: Open-Weight Agentic Coding Model Examined - Beagle · 2026-07-11Kimi K2.7 Code review: benchmarks, pricing, and the real cost | eesel AI · 2026-07-09Kimi K2.7-Code Review: Open-Weights 1T MoE Coding Model Deep Dive · 2026-06-20Kimi K2.7 Code: Moonshot's Coding-First Fork of a 1T MoE Base, Now with 30% Less Thinking — Redeep · 2026-06-12Weights 2026-06-11 · 2026-06-11Weights 2026-06-11now · 13 providers
JulAugSepOct2026-06-15: 1 repo commit2026-06-12: 1 repo commit2026-06-11: 2 repo commitsKimi K2.7 Code is generally available in GitHub Copilot · 2026-07-01Kimi K2.7 Code · 2026-06-13Kimi K2.7 Code · 2026-06-13Kimi-K2.7-Code · 2026-06-13Kimi K2.7 Code Review: Open Weights Enter Copilot | Awesome Agents · 2026-07-13Kimi K2.7-Code: Open-Weight Agentic Coding Model Examined - Beagle · 2026-07-11Kimi K2.7 Code review: benchmarks, pricing, and the real cost | eesel AI · 2026-07-09Kimi K2.7-Code Review: Open-Weights 1T MoE Coding Model Deep Dive · 2026-06-20Kimi K2.7 Code: Moonshot's Coding-First Fork of a 1T MoE Base, Now with 30% Less Thinking — Redeep · 2026-06-12Weights 2026-06-11 · 2026-06-11Weights 2026-06-11now · 13 providers

What people built on it

41 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

41 derivatives on the Hub, most of them quantized

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

28 quantized (Hugging Face model tree)11 fine-tunes (Hugging Face model tree)2 merges (Hugging Face model tree)28 quantized11 fine-tunes2 merges
28 quantized (Hugging Face model tree)11 fine-tunes (Hugging Face model tree)2 merges (Hugging Face model tree)28 quantized11 fine-tunes2 merges
GGUF unsloth/Kimi-K2.7-Code-GGUF for llama.cpp, Ollama, LM Studio 1.3M downloads MLX spicyneuron/Kimi-K2.7-Code-MLX-3.6bit for Apple silicon 12.8k downloads NVFP4 nvidia/Kimi-K2.7-Code-NVFP4 for vLLM, SGLang 1.2M downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

Benchmarks

6 rows

Benchmarks 6 rows

measured third-party leaderboard rowauthor-reported number from the model card or eval files (6)No confidence intervals unless the source gave them; settings differ between rows, so compare within a benchmark, not across. Date = the benchmark's own date when it gives one, else when we read it.
BenchmarkValueSetting / notesReported byDate
author-reported Kimi Claw 24/7 Bench46.9—model author2026-10-07
author-reported Kimi Code Bench v262—model author2026-10-07
author-reported MCP Atlas76—model author2026-10-07
author-reported MCP Mark Verified81.1—model author2026-10-07
author-reported MLS Bench Lite35.1—model author2026-10-07
author-reported Program Bench53.6—model author2026-10-07

Use-case evidence

7 use cases

What it is good at 7 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

22 fields

Listed in LLMs

Facts

Pipeline tagimage-text-to-text
Librarytransformers
ArchitecturesKimiK25ForConditionalGeneration
Model typekimi_k25
Dtypebfloat16
Params by dtypeBF16: 12,193,329,648 · F32: 23,040 · I32: 1,014,686,023,680
Size bandXXL
MoEyes
Experts384
Experts per token8
Context length262,144 (config.json max_position_embeddings)
LicenseKimi (Modified MIT) (modified-mit) (other)
Hub snapshot2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · KimiK25ForConditionalGeneration · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · KimiK25ForConditionalGeneration · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for Kimi-K2.7-Code: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)2.46 TB · Cluster (≥640 GB)
FP8 (1 byte/param)1.23 TB · Cluster (≥640 GB)
INT4 (0.5 bytes/param)616 GB · Multi-GPU node (<640 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra. MoE: all experts must be resident, so total parameters count.

All variants

40 variants

Lineage

Original release (shipped quantized by the lab) Family: moonshotai/Kimi-K2.7 41 derivatives on the Hub

quantized: 28finetune: 11merge: 2

Top variants (40 shown, ranked by downloads)

Relation
Quant
Size
40 rows
Variant Relation Quant Params Band License Author Likes Downloads
unsloth/Kimi-K2.7-Code-GGUF quantized GGUF — — other unsloth 197 1.3M
nvidia/Kimi-K2.7-Code-NVFP4 quantized NVFP4 — — other nvidia 19 1.2M
decart-ai/Kimi-K2.7-Code-NVFP4 quantized NVFP4 — — other decart-ai 2 414.9k
amd/Kimi-K2.7-Code-MXFP4 quantized MXFP4 518B XXL other amd 3 32.6k
eadx/Kimi-K2.7-Code-GGUF quantized GGUF — — other eadx 1 16.8k
novita/kimi-k2.7-code-eagle3-mla finetune — 1.84B S mit novita 3 16.1k
AesSedai/Kimi-K2.7-Code-GGUF quantized GGUF — — — AesSedai 9 13.1k
spicyneuron/Kimi-K2.7-Code-MLX-3.6bit quantized MLX 1.03T XXL — spicyneuron 9 12.8k
avlp12/Kimi-K2.7-Code-Alis-MLX-Dynamic-3.6bpw-VLM quantized MLX 1.03T XXL other avlp12 2 11k
mlx-community/Kimi-K2.7-Code-mlx-DQ3_K_M-q8 quantized mlx 1.03T XXL other mlx-community 5 10.5k
inferencerlabs/Kimi-K2.7-Code-MLX-Q3i quantized MLX — — — inferencerlabs 4 9.2k
DevQuasar/moonshotai.Kimi-K2.7-Code-GGUF quantized GGUF — — — DevQuasar 0 8.6k
Edmon02/Kimi-K2.7-Code-GGUF quantized GGUF — — other Edmon02 1 8.3k
CaelumAethryn/Kimi-K2.7-Code-GGUFcopy quantized GGUF — — other CaelumAethryn 0 7.5k
avlp12/Kimi-K2.7-Code-Alis-MLX-Dynamic-3.6bpw quantized MLX — — other avlp12 2 7.4k
pipenetwork/Kimi-K2.7-Code-MLX-4bit-hiprec quantized MLX 1.03T XXL other pipenetwork 1 6.2k
NullVoider/Kimi-K2.7-Code-GGUF quantized GGUF — — other NullVoider 0 5.3k
nvidia/Kimi-K2.7-Code-DFlash finetune — 3.48B M other nvidia 15 4.4k
SergeyKlevan/Kimi-K2.7-Code-GGUF quantized GGUF — — other SergeyKlevan 0 5.1k
aidiffuser/Kimi-K2.7-Code-vision finetune — — — other aidiffuser 2 4.9k
freakyskittle/kimi-k2.75-code-GGUF merge GGUF — — other freakyskittle 1 4.2k
aisoc/Kimi-K2.7-Code-vision finetune — — — other aisoc 0 4.2k
propelhq/Kimi-K2.7-Code-mlx-DQ3_K_M-q8 quantized mlx 1.03T XXL other propelhq 0 4k
bearzi/Kimi-K2.7-Code-JANGTQ_K finetune — 97.0B XL other bearzi 1 3.5k
freakyskittle/Kimi-K2.7-Code-Dflash quantized GGUF — — other freakyskittle 6 2.5k
Chen2921/Kimi-K2.7-Code-GGUF quantized GGUF — — — Chen2921 0 2.7k
unsloth/Kimi-K2.7-Code quantized — 1.03T XXL other unsloth 22 1.6k
cm00cm/Kimi-K2.7-Code-DFlash finetune — 3.48B M other cm00cm 1 2.4k
mradermacher/Kimi-K2.7-Code-i1-GGUF quantized GGUF — — other mradermacher 1 1.8k
novita/kimi-k2.7-code-dspark finetune — 3.15B M — novita 7 1.3k
freakyskittle/kimi-k2.75-code merge — 742B XXL other freakyskittle 1 1.5k
cm00cm/Kimi-K2.7-Code-EAGLE3 finetune — 1.45B S other cm00cm 1 775
mradermacher/Kimi-K2.7-Code-GGUF quantized GGUF — — other mradermacher 2 684
nikita-savelyev-cerebras/tiny-random-kimi-k2.7-code-int4 finetune int4 3.24M S — nikita-savelyev-cerebras 0 501
batiai/Kimi-K2.7-Code-GGUF quantized GGUF — — other batiai 0 371
gghfexp/Kimi-K2.7-Code-GGUF quantized GGUF — — other gghfexp 1 289
loxxxxx/Huihui-Kimi-K2.7-Code-abliterated-GGUF quantized GGUF — — other loxxxxx 2 182
mixlayer/Kimi-K2.7-Code-0.7B-A0.4B finetune — 678M S mit mixlayer 0 142
rAVEUK/Kimi-K2.7-Code quantized — 1.03T XXL other rAVEUK 0 61
PYTHAI/Kimi-K2.7-Code-fork finetune — — — other PYTHAI 0 35

Repository changelog

4 commits
changed in the last 7 daysRank band moves: 2026-10-23all changes

Repository changelog 4 commits

DateCommitSHA
2026-06-15 force thinking & preserve_thinking in jinja; set default sampling params in generation_config.json 74797c9
2026-06-12 Update README.md 78b5c46
2026-06-11 Add files using upload-large-folder tool da26fe4
2026-06-11 initial commit fcc7483

All providers

13 providers · 13 endpoints

Where to run 13 providers · 13 endpoints

Hosted endpoints observed for Kimi-K2.7-Code, sorted by input price (USD per million tokens). Each row names its observer and date; prices change daily, so check the source before quoting. Cheapest observed input price: $0.67/M.

ProviderModel refContextQuant$/M in$/M outtok/sTTFTToolsStatusObserved
Inceptron
openrouter
 moonshotai/kimi-k2.7-code 256K int4 $0.67 $3.35 — — yes live 2026-10-24
DeepInfra
hf
 moonshotai/Kimi-K2.7-Code 256K — $0.68 $3.40 34.5 625 ms yes live 2026-10-25
CoreWeave
openrouter
 moonshotai/kimi-k2.7-code 256K int4 $0.71 $3.50 — — yes live 2026-10-24
StreamLake
openrouter
 moonshotai/kimi-k2.7-code 250K unknown $0.71 $3.00 — — yes live 2026-10-24
Venice
openrouter
 moonshotai/kimi-k2.7-code 250K int4 $0.75 $3.50 — — yes live 2026-10-24
SiliconFlow
openrouter
 moonshotai/kimi-k2.7-code 256K fp8 $0.86 $3.80 — — yes live 2026-10-24
Novita
hf, openrouter
 moonshotai/kimi-k2.7-code 256K int4 $0.91 $3.84 73.4 1643 ms yes live 2026-10-25
Nebius
openrouter
 moonshotai/kimi-k2.7-code 256K fp4 $0.95 $4.00 — — yes live 2026-10-24
GMICloud
openrouter
 moonshotai/kimi-k2.7-code 256K fp8 $0.95 $4.00 — — yes live 2026-10-24
Alibaba
openrouter
 moonshotai/kimi-k2.7-code 256K fp8 $0.95 $4.00 — — yes live 2026-10-24
Cloudflare
openrouter
 moonshotai/kimi-k2.7-code 256K unknown $0.95 $4.00 — — yes live 2026-10-24
Moonshot AI
openrouter
 moonshotai/kimi-k2.7-code 256K int4 $0.95 $4.00 — — yes live 2026-10-24
Featherless
hf
 moonshotai/Kimi-K2.7-Code — — — — — — — live 2026-10-25

Articles and coverage

12 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Official3

Kimi K2.7 Code
kimi.com2026-06-13officialauthorityHN 4 pts · 1 comments539 words
Kimi K2.7 Code is an open-source programming model from Moonshot AI featuring agent capabilities and a 1T parameter MoE architecture with 32B activated parameters per token. The model supports a 256K context length and reduces thinking token usage by approximately 30% compared to K2.6 while achieving higher scores on Kimi Code Bench v2 and Program Bench. API pricing is ¥1.30 for cached input, ¥6.50 for uncached input, and ¥27.00 for output per 1M tokens.written, reviewed, cited
Kimi K2.7 Code
platform.kimi.ai2026-06-13officialHN 4 pts · 1 comments753 words
Kimi K2.7 Code is a dedicated coding model that significantly improves instruction compliance and long-horizon coding performance compared to K2.6. The model reduces overthinking tendencies by 30% on average and offers a 10% performance improvement in Agentic capabilities. It supports a 256K context window, lacks non-thinking mode support, and includes a high-speed variant reaching up to 260 Tokens/s.written, reviewed, cited
  • The Vision model has no limit on the number of images, but the request body size must not exceed 100M.
  • For very large videos, the file upload method is mandatory because URL-formatted images are not supported.
  • When using tools, tool_choice can only be set to 'auto' or 'none' to avoid conflicts with reasoning content.
Kimi K2.7 Code: Open-Source Agentic Coding Model - Kimi.ai
kimi.aiofficial1,108 words
Moonshot AI released Kimi K2.7 Code, an open-source coding model built on a Mixture-of-Experts architecture with 1 trillion total parameters and 32 billion activated parameters. The model supports a 256K context length and reduces thinking-token usage by approximately 30% compared to K2.6 while achieving higher scores on coding and agentic benchmarks. Kimi K2.7 Code is available via the Kimi API with input pricing of $0.19 per million tokens for cache hits and $0.95 for cache misses.written, reviewed, cited
  • Kimi K2.7 Code does not support non-thinking mode and always runs with thinking enabled.
  • Requests made with thinking disabled in Kimi Code are automatically served by the K2.6 model instead.
  • The model includes MoonViT, a 400M-parameter vision encoder integrated into its architecture.

Analysis6

Kimi K2.7 Code is generally available in GitHub Copilot
github.blog2026-07-01authorityHN front pageHN 417 pts · 185 comments317 words
Kimi K2.7 Code is generally available in GitHub Copilot as the first open-weight model offered as a selectable option in the Copilot model picker. The model is beginning to roll out to Copilot Pro, Pro+, and Max plans within Visual Studio Code version 1.127.0 or later, while it remains off by default for Copilot Business and Copilot Enterprise requiring plan administrators to enable the policy. Rollout will be gradual as GitHub continues to monitor the model's quality and performance.written, reviewed, cited
Kimi K2.7 Code Review: Open Weights Enter Copilot | Awesome Agents
awesomeagents.ai2026-07-131,982 words
Kimi K2.7-Code: Open-Weight Agentic Coding Model Examined - Beagle
heybeagle.com2026-07-11
Kimi K2.7 Code review: benchmarks, pricing, and the real cost | eesel AI
eesel.ai2026-07-092,159 words
Kimi K2.7-Code Review: Open-Weights 1T MoE Coding Model Deep Dive
open-techstack.com2026-06-202,366 words
Kimi K2.7 Code: Moonshot's Coding-First Fork of a 1T MoE Base, Now with 30% Less Thinking — Redeep
redeep.blog2026-06-121,510 words

Community1

Kimi-K2.7-Code
twitter.com2026-06-13HN 5 pts · 1 comments
All coverage tracked for this model (12) — append-only, nothing is ever removed
TitleSourcePublishedFirst seenSignals
Kimi K2.7 Code: Open-Source Agentic Coding Model - Kimi.aikimi.ai official—2026-10-07exa
Kimi K2.7-Code cuts thinking tokens 30%venturebeat.com—2026-10-07exa
Kimi K2.7-Code is Out. The Open-Source Coding Model That Thinks Lessdishantsharma.hashnode.dev—2026-10-07exa
Kimi K2.7 Code Review: Open Weights Enter Copilot | Awesome Agentsawesomeagents.ai2026-07-132026-10-07exa
Kimi K2.7-Code: Open-Weight Agentic Coding Model Examined - Beagleheybeagle.com2026-07-112026-10-07exa
Kimi K2.7 Code review: benchmarks, pricing, and the real cost | eesel AIeesel.ai2026-07-092026-10-07exa
Kimi K2.7 Code is generally available in GitHub Copilotgithub.blog2026-07-012026-10-07hn
Kimi K2.7-Code Review: Open-Weights 1T MoE Coding Model Deep Diveopen-techstack.com2026-06-202026-10-07exa
Kimi K2.7 Codekimi.com official2026-06-132026-10-07hn
Kimi K2.7 Codeplatform.kimi.ai official2026-06-132026-10-07hn
Kimi-K2.7-Codetwitter.com2026-06-132026-10-07hn
Kimi K2.7 Code: Moonshot's Coding-First Fork of a 1T MoE Base, Now with 30% Less Thinking — Redeepredeep.blog2026-06-122026-10-07exa

Similar models

3 models

Similar models

  • Kimi-K2.6 · 1.03T
    same size band (XXL), also MoE, same task (image-text-to-text), more providers
  • Kimi-K2.5 · 1.03T
    same size band (XXL), also MoE, same task (image-text-to-text)
  • Inkling · 952B
    same size band (XXL), also MoE, same task (image-text-to-text), permissive license

Sources

4 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:34 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:45 UTC · Summary: 2026-10-07 15:01 UTC · Last verified 2026-10-25

Follow Kimi-K2.7-Code

Watch this model

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.