Kimi releases K3.1 Pro, a 3T-class open-weight model with 1M-token context and faster highspeed API

𝕏/@kimi_moonshot •

Revision history

3 recorded changes

Want your article here?

Kimi says K3.1 Pro is a 3T-class open-weight multimodal MoE model with 1M-token context, 2.8T total and 104B active parameters. It scores 65.7% on Terminal-Bench 4.0, ahead of GPT-6 Astra (58.2%) and just behind Claude Opus 5.5 (66.4%), and 92.7% on BenchCAD. The company reports hallucinations down 65% on finance work, plus gains over K3 on financial modeling and research. A 2–3× faster Highspeed variant is live; open weights are promised by Nov 4, 2026.

AI summary of the linked source — not editorial content.

Story signal · how it spread on X · via Elfa

as of 23 Oct 18:04 UTC

Related on X Moonshot limits Kimi K3.1 Pro’s open-weight cyber capabilities before release · 2 accounts · first post 23 Oct 17:34 UTC

Kimi releases K3.1 Pro, a 3T-class open-weight model with 1M-token context and faster highspeed API

Story signal · how it spread on X · via Elfa

Explore the topic

More on Kimi

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Total stats

How to Earn

More on Kimi

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Kimi releases K3.1 Pro, a 3T-class open-weight model with 1M-token context and faster highspeed API

Story signal · how it spread on X · via Elfa

Explore the topic

More on Kimi

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Comments

Total stats

How to Earn

More on Kimi

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API

Moonshot teases Kimi K 3.1 release

Moonshot’s Kimi faces fresh scrutiny after U.S. agencies allege Chinese AI firms systematically extracted knowledge from American frontier models at industrial scale

0G added Moonshot's Kimi K3 to its decentralized AI routing network, exposing the Chinese frontier model through a hardware-attested path despite mounting scrutiny from Washington

Moonshot's new Kimi K3 is the largest open weight model ever at 2.8T param, and scores the same as GPT-5.6 and Fable 5 on benchmarks. However, it is signficantly cheaper at $3/15 per M input/output, Sonnet 5 pricing

Kimi K3 rumors and ArenaML leaks point to imminent release

Cursor Composer 2 outed as fine-tuned Kimi K2.5 after developer spots internal model ID leaking through API