DeepSWE
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
Context window on the 2026-08-19 OpenRouter row: 262,144 tokens.
Max completion tokens on that row: 100,352.
Input / output on that row: $0.60 / $2.50 per 1M tokens.
Architecture fields: text->text · Other.
Hugging Face id on that row: moonshotai/Kimi-K2-Thinking.
Kimi K2 Thinking is Moonshot AI’s most advanced open reasoning model to date, extending the K2 series into agentic, long-horizon reasoning.
When you want the 2025-11-06 catalog SKU, not a later rename.
No official DeepSWE mini-swe-agent row for this identity on the public board we fetched.
We did not open a vals.ai card for this identity. Chip omitted.
Ranked view: Best coding models: the independent leaderboard puts this row and every other card that carries an independent coding score on one board, with the confidence intervals left visible.
Copied from AA or vals HTML we opened. 0.0% placeholder bars are omitted. Do not average these into the hero tiles, and do not invent CI, $/task, or steps for Artificial Analysis.
| Bench | Printed | Note | Source | URL | As of |
|---|---|---|---|---|---|
| SciCode | 42.4% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
| Humanity's Last Exam | 23.8% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
| GPQA Diamond | 83.8% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
| CritPt | 2.6% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
| AA-Omniscience Accuracy | 30.9% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
| AA-LCR | 70.3% | Printed on the AA Intelligence Evaluations grid for Kimi K2 Thinking. Not a DeepSWE chip. | Artificial Analysis Kimi K2 Thinking | source | 2026-08-19 |
Plus is $25/mo with $25 weekly hosted usage. The Mac app stays free with your own keys. Get Plus is not Download for Mac.
Official weights id on the 2026-08-19 OpenRouter row: moonshotai/Kimi-K2-Thinking. Community GGUF is a quant, not a second Continuum card.
This is the model repo, not Download for Mac. Downloads and likes from the Hugging Face API on 2026-08-19.
usedStorage and the sum of .safetensors sizes can differ (LFS pointers, non-shard files, duplicate copies). Cited separately. Source: recursive tree API, 2026-08-19.
| Path | Bytes |
|---|---|
model-00001-of-000062.safetensors | 995,002,080 |
model-00002-of-000062.safetensors | 9,808,995,784 |
model-00003-of-000062.safetensors | 9,808,995,784 |
model-00004-of-000062.safetensors | 9,808,995,784 |
model-00005-of-000062.safetensors | 9,808,995,784 |
model-00006-of-000062.safetensors | 9,808,995,784 |
model-00007-of-000062.safetensors | 9,808,995,784 |
model-00008-of-000062.safetensors | 9,808,995,784 |
model-00009-of-000062.safetensors | 9,808,995,784 |
model-00010-of-000062.safetensors | 9,808,995,784 |
model-00011-of-000062.safetensors | 9,808,999,256 |
model-00012-of-000062.safetensors | 9,808,999,256 |
model-00013-of-000062.safetensors | 9,808,999,256 |
model-00014-of-000062.safetensors | 9,808,999,256 |
model-00015-of-000062.safetensors | 9,808,999,256 |
model-00016-of-000062.safetensors | 9,808,999,256 |
model-00017-of-000062.safetensors | 9,808,999,256 |
model-00018-of-000062.safetensors | 9,808,999,256 |
model-00019-of-000062.safetensors | 9,808,999,256 |
model-00020-of-000062.safetensors | 9,808,999,256 |
model-00021-of-000062.safetensors | 9,808,999,256 |
model-00022-of-000062.safetensors | 9,808,999,256 |
model-00023-of-000062.safetensors | 9,808,999,256 |
model-00024-of-000062.safetensors | 9,808,999,256 |
model-00025-of-000062.safetensors | 9,808,999,256 |
model-00026-of-000062.safetensors | 9,808,999,256 |
model-00027-of-000062.safetensors | 9,808,999,256 |
model-00028-of-000062.safetensors | 9,808,999,256 |
model-00029-of-000062.safetensors | 9,808,999,256 |
model-00030-of-000062.safetensors | 9,808,999,256 |
model-00031-of-000062.safetensors | 9,808,999,256 |
model-00032-of-000062.safetensors | 9,808,999,256 |
model-00033-of-000062.safetensors | 9,808,999,256 |
model-00034-of-000062.safetensors | 9,808,999,256 |
model-00035-of-000062.safetensors | 9,808,999,256 |
model-00036-of-000062.safetensors | 9,808,999,256 |
model-00037-of-000062.safetensors | 9,808,999,256 |
model-00038-of-000062.safetensors | 9,808,999,256 |
model-00039-of-000062.safetensors | 9,808,999,256 |
model-00040-of-000062.safetensors | 9,808,999,256 |
model-00041-of-000062.safetensors | 9,808,999,256 |
model-00042-of-000062.safetensors | 9,808,999,256 |
model-00043-of-000062.safetensors | 9,808,999,256 |
model-00044-of-000062.safetensors | 9,808,999,256 |
model-00045-of-000062.safetensors | 9,808,999,256 |
model-00046-of-000062.safetensors | 9,808,999,256 |
model-00047-of-000062.safetensors | 9,808,999,256 |
model-00048-of-000062.safetensors | 9,808,999,256 |
model-00049-of-000062.safetensors | 9,808,999,256 |
model-00050-of-000062.safetensors | 9,808,999,256 |
model-00051-of-000062.safetensors | 9,808,999,256 |
model-00052-of-000062.safetensors | 9,808,999,256 |
model-00053-of-000062.safetensors | 9,808,999,256 |
model-00054-of-000062.safetensors | 9,808,999,256 |
model-00055-of-000062.safetensors | 9,808,999,256 |
model-00056-of-000062.safetensors | 9,808,999,256 |
model-00057-of-000062.safetensors | 9,808,999,256 |
model-00058-of-000062.safetensors | 9,808,999,256 |
model-00059-of-000062.safetensors | 9,808,999,256 |
model-00060-of-000062.safetensors | 9,808,999,256 |
model-00061-of-000062.safetensors | 9,808,999,256 |
model-00062-of-000062.safetensors | 4,697,635,112 |
model.safetensors.index.json | 20,450,783 |
tiktoken.model | 2,795,286 |
figures/banner.png | 291,736 |
figures/Base-Evaluation.png | 245,449 |
figures/kimi-logo.png | 87,988 |
modeling_deepseek.py | 75,769 |
README.md | 16,176 |
tokenization_kimi.py | 12,586 |
docs/tool_call_guidance.md | 11,338 |
configuration_deepseek.py | 10,652 |
docs/deploy_guidance.md | 4,172 |
tokenizer_config.json | 4,046 |
config.json | 3,830 |
chat_template.jinja | 3,450 |
.gitattributes | 1,849 |
THIRD_PARTY_NOTICES.md | 1,673 |
LICENSE | 1,462 |
generation_config.json | 53 |
All 80 files the tree API returned are listed: every one of the 62 safetensor shards plus 18 other files. Files tab: https://huggingface.co/moonshotai/Kimi-K2-Thinking/tree/main.
OpenRouter slug moonshotai/kimi-k2-thinking on the 2026-08-19 catalog.
First party: https://platform.moonshot.ai/docs/overview.
Not on the Continuum host list fetched 2026-08-19.
We list first-party and OpenRouter, plus Continuum only when the live public allowlist named the id. This is not a 15-host routing table.
OpenRouter 2026-08-19: $0.60 in / $2.50 out per 1M tokens. Context 262,144 in / 100,352 out.
No separate internal-reasoning price on that row.
| Claim | Source | As of |
|---|---|---|
| OpenRouter id moonshotai/kimi-k2-thinking; context 262144; created 1762440622; $0.60 / $2.50 per 1M | OpenRouter /api/v1/models | 2026-08-19 |
| AA Intelligence Index 33; 6 printed Index benches | Artificial Analysis model page | 2026-08-19 |
| OpenRouter id moonshotai/kimi-k2-thinking, context 262,144 | OpenRouter /api/v1/models | 2026-08-19 |
| HF downloads 53,004 | Hugging Face API moonshotai/Kimi-K2-Thinking | 2026-08-19 |
| HF safetensors.total: 1,058,118,284,416 stored tensors; HF API usedStorage 594,283,168,582 bytes; created 2025-11-04T08:25:31.000Z; HF tensors F32 + I32 + BF16 | Hugging Face API moonshotai/Kimi-K2-Thinking | 2026-08-19 |
| 62 safetensor shards; shard bytes 594,232,561,304; tree files 80 | Hugging Face tree API moonshotai/Kimi-K2-Thinking | 2026-08-19 |
No other card in this cluster shares a published DeepSWE official row we can put next to this one. We will not compare on vendor-blog numbers.
Continue through Moonshot's model family with Kimi K2 0905.
Lab posts pick a harness, an effort, a split, and sometimes a private eval set. DeepSWE publishes the official mini-swe-agent row with cost, tokens, and steps. Scale SWE-bench Pro only counts when the public shared-harness board has a row we can fetch. A higher lab number is usually a different test, not a better one.
No Continuum hosted id is verified for this model, so this page does not invent a curl target. Use the first-party API or the OpenRouter slug moonshotai/kimi-k2-thinking.