AI models

The frontier and open-weights models developers are building on — with the specs that actually matter: context window, modalities, licensing and price per token.

Mercury 3

Inception

A diffusion-based language model for reasoning, tool use, and latency-sensitive agentic workflows, served through Inception's API.

API1M ctx · 128K outReasoningTools
$0.1 / $0.35 per 1M
5w ago

Muse Spark Open

Meta

An open-weight, natively multimodal model for coding, long-horizon agentic work, and complex software engineering.

Open weights1M ctx · 944K out816B total / 42B active (MoE)ReasoningToolsVisionAudioPDF

Nemotron 3.5 Super VL

NVIDIA

An open-weight multimodal mixture-of-experts model for specialized agents, visual reasoning, coding, and tool use.

Open weights1M ctx · 8K out120B total / 12B active (MoE)ReasoningToolsVision

Gemini 4 Argon

Google DeepMind

Gemini 4 Argon is Google's frontier reasoning model for complex software engineering, mathematical and scientific problem solving, and long-running agentic workflows.

API1M ctx · 32K outReasoningToolsVisionAudioPDF
$2 / $10 per 1M
5w ago

Muse Wonder

Meta

A frontier model for personal and professional agents, combining Active Memory and Active Focus for sustained coding and complex tasks. Available through the Meta Model API and Muse Code.

API1M ctx · 128K outReasoningToolsVisionAudioPDF
$4 / $20 per 1M
6w ago

Mistral Large 4

beta

Mistral AI

Mistral Large 4 is a natively multimodal model with 1 trillion total parameters and 49 billion active parameters, built for coding, agentic workflows and multimodal understanding. Available in public API preview; weights are planned for release by the end of October 2026.

API512K ctx · 256K outReasoningToolsVision
$0.68 / $2.09 per 1M
6w ago

flow-1

Laminar

flow-1 is our new model for understanding agent traces, trained with reinforcement learning to identify hard-to-spot failures and explain why they happened. On our benchmark, it matches GPT-6-sol in detection quality while being 23x cheaper.

APITools
7w ago

Clef

Cloudflare

Clef is a 27B multimodal model that turns a state and a schema of typed questions into decisions. It reads the state as text, JSON, images, or video, and returns a probability for every allowed option of every question in a single forward pass. There is no free-form text generation and no output parsing.

Open weights27B denseVision

GPT-6.1 Sol

OpenAI

Near-Astra performance for complex coding, computer use, and professional work at a lower cost.

API1.05M ctx · 128K outReasoningToolsVisionPDF
$2 / $10 per 1M
7w ago

Claude Sonnet 5.5

Anthropic

Claude Sonnet 5.5 is a clear upgrade over Claude Sonnet 5, runs 30%+ faster, and costs up to 30% less for most work.

API1M ctx · 128K outReasoningToolsVisionPDF
$2 / $10 per 1M
8w ago

Gemini 3.8 Flash TTS

Google

Built for deep creative direction and character design. Create entirely new voices from scratch using natural language prompts to bring characters to life across gaming, immersive audiobooks, podcasts, and interactive media. Direct every performance line by line with granular control over acting cues, pacing, dialect shifts, and backchanneling.

APIAudio out
8w ago

Gemini 3.8 Flash-Lite TTS

Google

Built for high-volume, cost-efficient scale. Optimized for high-volume dubbing, audio content creation, and expressive voice agents with fine-grained control over tone, pacing, and expressive nuance.

APIAudio out
8w ago

Claude Opus 5.5

Anthropic

Claude Opus 5.5 leads in agentic coding and knowledge work, and costs 40% less to run than Opus 5 on typical workloads.

API1M ctx · 128K outReasoningToolsVisionPDF
$4 / $20 per 1M
8w ago

FLUX 3 Action

Black Forest Labs

FLUX 3 Action is an open weights 7B world action model. It takes camera frames, the robot's state and a text instruction, and returns the next chunk of actions, denoised together with the next video frames.

Open weights7BVision

GPT-6 Luna

OpenAI

OpenAI's most efficient model for focused, high-volume tasks

API1.05M ctx · 128K outReasoningToolsVisionPDF
$0.1 / $0.5 per 1M
8w ago

GPT-6 Sol

OpenAI

OpenAI model for complex coding and agentic workflows

API1.05M ctx · 128K outReasoningToolsVisionPDF
$2 / $10 per 1M
8w ago

Ternary Bonsai 2 27B

PrismML

A ternary 27B multimodal reasoning model based on Qwen3.8, supporting coding, mathematics, image understanding, and tool use.

Open weights256K ctx · 32K out27BReasoningToolsVision

Needle 3

Cactus Compute

A 121M-parameter open model for tool calling, structured extraction and text embedding on tiny devices. One set of weights ships as an intelligence ladder: every depth from 2 to 20 layers is a deployable model, an 8-29 MB CQ2-bit binary.

Open weights8K ctx121MTools

Gemini 3.8 Live

Google

Built for scale and cost efficiency, combining conversational intelligence with fluid dialogue and visual grounding.

APIToolsAudioAudio out
9w ago

Gemini 3.8 Live Extended Thinking

Google

Built for high-complexity tasks, with increased intelligence and multi-step reasoning.

APIToolsAudioAudio out
9w ago

DeepSeek V4.1 Flash

DeepSeek

We introduce DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts (MoE) model with 552B backbone parameters and support for contexts of up to one million tokens. The model natively processes images and text, and generates text autoregressively.

Open weights1M ctx · 384K outReasoningToolsVision

GPT-Image-2.5 Flare

OpenAI

GPT-Image-2.5 Flare brings the same improvements in quality, editing, and speed.

APIVisionImage out
10w ago

GPT-Image-2.5 Sunburst

OpenAI

GPT-Image-2.5 Sunburst offers an extra level of precision for detailed creative work with longer generation times.

APIVisionImage out
10w ago

GPT-6 Astra

OpenAI

Our most capable model, built for the hardest end-to-end work.

API1.05M ctx · 128K outReasoningToolsVision
$10 / $50 per 1M
11w ago

MAI-Transcribe-2

Microsoft AI

Turn noisy audio into precise, domain-specific transcripts, with leading FLEURS and Artificial Analysis accuracy scores.

APIAudio
11w ago

Gemini 3.8 Flash

Google

Our most intelligent workhorse model, delivering significant improvements from 3.7 Flash across software engineering, agentic tasks, and critical, multi-step reasoning in specialized domains.

API1M ctx · 64K outReasoningToolsVisionAudioPDF
$0.75 / $3.75 per 1M
11w ago

Gemini 3.8 Flash Cyber

Google

Our most capable cybersecurity model with frontier-level performance in vulnerability detection and automated patching, available to trusted defenders through our new Fairwind Program.

API
11w ago

Muse Spark 1.3

Meta

A multimodal reasoning model for coding and long-running agentic workflows, improving instruction following and efficiency over Muse Spark 1.2.

API1M ctx · 128K outReasoningToolsVisionAudioPDF
$1.25 / $4.25 per 1M
11w ago

Raptor 0.5

Osaurus

An 8B mixture-of-experts that activates only ~1B parameters per token, runs the full Osaurus tool surface in 6.3 GB, and stays resident on the 8–16 GB Macs most people own.

Open weights128K ctx7.9B total, ~1B active (MoE)ReasoningTools

Atlas

World Labs

Atlas is an omni model that we pretrained from scratch to natively operate on text, images, video, and 3D. It is a multimodal autoregressive diffusion transformer: all inputs are combined into a shared spatial context.

APIVisionImage outVideo out
11w ago

Claude Fable 5.1

Anthropic

Claude Fable 5.1 sets a new standard for coding, knowledge work, and long-running problem-solving tasks.

API1M ctx · 128K outReasoningToolsVisionPDF
$10 / $50 per 1M
11w ago

Claude Mythos 5.1

Anthropic

Mythos 5.1 is available only through our trusted access programs; its safeguards are specifically designed to support work in cybersecurity and the life sciences.

API1M ctx · 128K outReasoningToolsVisionPDF
$10 / $50 per 1M
11w ago

Gemini Omni 1.1 Flash

Google

Omni now delivers studio-quality video production, including the ability to extend a scene, first and last frame interpolation, crisp 4K upscaling, faster prototyping, and more.

APIVisionVideo out
12w ago

Gemini 3.5 Transcribe

Google

Our most precise speech-to-text model yet, designed for intelligent voice interactions.

APIToolsAudio
$2 / $12 per 1M
12w ago

Gemini 3.5 Transcribe Live

Google

Our most precise speech-to-text model yet, designed for intelligent voice interactions.

APIToolsAudio
$3.5 / $21 per 1M
12w ago

GLM-5.3-Flash

Zhipu AI

GLM-5.3-Flash is the first native multimodal model in the GLM-5 series, delivering stronger intelligence than GLM-5.2 while maintaining an exceptionally cost-efficient architecture.

Open weights1M ctx · 128K out320B total, 18B activeReasoningToolsVisionPDF

GLiNER2.5 Base

Fastino Labs

A new boundary-prediction architecture replaces span enumeration, adding joint entity-relation extraction, constrained classification, and unlimited span length.

Open weights

GLiNER2.5 Multilingual

Fastino Labs

A new boundary-prediction architecture replaces span enumeration, adding joint entity-relation extraction, constrained classification, and unlimited span length.

Open weights

GLiNER2.5 Small

Fastino Labs

A new boundary-prediction architecture replaces span enumeration, adding joint entity-relation extraction, constrained classification, and unlimited span length.

Open weights

Qwen3.8 Flash Next

Qwen

An open-weights multimodal MoE model that doubles as an early preview of the Qwen4 architecture, the same role Qwen3-Next played for Qwen3.5. It pairs Gated DeltaNet with Qwen Sparse Attention, widens the residual stream into four gated branches, and adds 51B N-gram embedding parameters that cost almost nothing per token. Natively 262K context, extensible to 1M with YaRN.

Open weights256K ctx125B total, 6B active, plus 51B N-gram embeddingsReasoningToolsVision

GLM-5.3

Zhipu AI

GLM-5.3 is the most capable open-weights model for coding, with a 50% improvement over GLM-5.2 on our in-house Z.ai Code Bench.

API1M ctx · 128K outReasoningTools
$1.4 / $4.4 per 1M
14 Aug 2026

Qwen3.8 27B

Qwen

Built on the architectural foundation of Qwen3.5, Qwen3.8 delivers substantial gains across coding, professional work, research, and long-horizon agentic tasks. Qwen3.8-27B brings these advances to a compact, deployment-friendly dense model: a native vision-language model that understands images and videos, with flexible thinking control, designed to carry complex, multi-step tasks through to completion with greater reliability.

Open weights256K ctx · 128K out27B denseReasoningToolsVision

Gemini 3.7 Flash

Google

Our most intelligent workhorse model yet for coding and agents.

API1M ctx · 64K outReasoningToolsVisionAudioPDF
$0.75 / $3.75 per 1M
13 Aug 2026

Grok 4.6

xAI

xAI's frontier Grok model, tuned for long-running agents, coding, knowledge work, and visual projects.

API500K ctx · 500K outReasoningToolsVisionPDF
$2 / $6 per 1M
12 Aug 2026

LFM2.5-VL-3B

Liquid AI

LFM2.5-VL-3B is the multimodal variant of LFM2.5, pairing the LFM2.5-2.6B backbone with a SigLIP2 NaFlex 400M vision encoder. It answers directly rather than reasoning, and adds screen understanding, grounding and function calling for on-device use.

Open weights32K ctx3.1BToolsVision
↓ 26.4K♥ 214
12 Aug 2026

S1-mini

Superwhisper

A 0.6B-parameter text normalizer for speech-to-text output. It takes a raw ASR transcript and rewrites it as clean written text: fillers removed, false starts and self-corrections resolved to the value the speaker landed on, punctuation and capitalization applied, and spoken numbers, dates, times, currency and email addresses rendered in written form.

Open weights40K ctx596M
↓ 4K♥ 360
12 Aug 2026

LTX-2.5

LTX

LTX-2.5 generates multi-shot scenes in one pass, edits real footage, and exports cinema-grade EXR. Open weights you can fine-tune and run on your hardware.

Open weightsVisionAudioVideo outAudio out
↓ 1.6M♥ 5.3K
11 Aug 2026

Needle 2

Cactus Compute

An open 45M-parameter model for tool calling, device use, and structured extraction. Needle 2 runs as a 14 MB binary in 28 MB of session RAM.

Open weights256 ctx45MTools
↓ 24K♥ 328
11 Aug 2026

Nemotron 3.5 Lightning

NVIDIA

A customizable open 30B MoE model with 3B active parameters, providing optimal high-volume execution for autonomous agents.

Open weights256K ctx · 256K out30B total, 3B active (MoE)ReasoningTools

Ling 3.0 Tiny

inclusionAI

We are introducing Ling-3.0-tiny, a lightweight hybrid reasoning MoE model with 7.9B total parameters and only 1.3B activated parameters per token. It is designed to deliver strong reasoning and agentic capabilities at low inference cost, making advanced model capabilities more accessible for local and resource-constrained deployment.

Open weights256K ctx · 32K out7.9B total, 1.3B active (MoE)ReasoningTools
↓ 18.9K♥ 491
10 Aug 2026

Muse Glimmer 30B

Meta

Muse Glimmer is a 30-billion-parameter causal language model with a dedicated perception encoder, distilled from Muse Spark and purpose-built for autonomous agentic tasks on consumer hardware. The model integrates multi-step reasoning, reliable tool use, multimodal understanding, and failure recovery into a single model that runs locally without requiring cloud infrastructure or network access.

Open weights128K ctx29.6BReasoningToolsVision
↓ 286.7K♥ 2K
10 Aug 2026

TwiL-LM 1.7B

webAI

A LoRA adapter over SmolLM2-1.7B-Instruct for formal logic — FOL translation, entailment, semantic parsing and Lean assistance — quantized to a 1.06 GB download that runs on a phone.

Open weights8K ctx1.7B
↓ 1.2K♥ 16
7 Aug 2026

TwiL-LM3

webAI

A 3B formal-logic reasoning model built from SmolLM3-3B by LoRA fine-tuning, checkpoint fusion, WiSE-FT interpolation and entropy-weighted GRPO, which beats gpt-oss-120b on four of five lanes of webAI's formal-reasoning suite.

Open weights64K ctx3BReasoning

FLUX 3

Black Forest Labs

FLUX 3 is our new multimodal foundation model. It jointly learns from images, videos, and audio within a unified architecture, because what it needs to learn is not any one of these elements in isolation. Instead, a model must learn a representation of the world: how objects hold together, how things move, and how events sound.

APIVisionAudioVideo outAudio out
4 Aug 2026

LFM2.5-2.6B

Liquid AI

LFM2.5-2.6B is part of LFM2.5, a family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with a 128K context window and agentic post-training.

Open weights128K ctx2.6BTools

Shieldstral

Mistral AI

A 3B open-weights, policy-adaptive multimodal safety classifier that matches models up to 7x its size on text safety and sets a new state of the art on multimodal moderation.

Open weights256K ctx3BVision

MiniMax H3

MiniMax

MiniMax H3 is a general-purpose, omni-modal generative system. It supports unified understanding of multimodal contexts composed of text, images, video, and audio, and can generate video with native stereo audio at resolutions up to 2K and durations of up to 15 seconds.

Open weights33BVisionAudioVideo outAudio out

Inkling-Small

Thinking Machines Lab

A quarter the size of Inkling at comparable performance: a multimodal Mixture-of-Experts reasoner (276B total, 12B active) with controllable reasoning effort, fine-tunable on Tinker.

Open weights

Claude Opus 5

Anthropic

Opus 5 is a step change improvement for the Opus tier powering long-running agents while delivering improvements in coding and professional work.

API1M ctx · 128K outReasoningToolsVisionPDF
$5 / $25 per 1M
24 Jul 2026

Gemini 3.5 Flash-Lite

Google

Our fastest, most cost-effective 3.5-class model, delivering 350 output tokens per second.

API1M ctx · 64K outReasoningToolsVisionAudioPDF
$0.3 / $2.5 per 1M
21 Jul 2026

Gemini 3.6 Flash

Google

Our workhorse model that delivers better coding, knowledge work, and multimodal performance.

API1M ctx · 64K outReasoningToolsVisionAudioPDF
$0.75 / $3.75 per 1M
21 Jul 2026

Laguna S 2.1

Poolside

The most capable agentic coding model in its weight class by a wide margin.

Open weights1M ctx · 32K out118B total, ~8B active (MoE)ReasoningTools
↓ 104.3K♥ 1K
21 Jul 2026

Kimi K3

Moonshot AI

Kimi K3 is a 2.8T-parameter model built on Kimi Delta Attention and Attention Residuals, with native vision capabilities and a 1-million-token context window — the world's first open 3T-class model.

Open weights1M ctx · 128K outReasoningToolsVision

Inkling

Thinking Machines Lab

Thinking Machines Lab's first open-weights model: a multimodal Mixture-of-Experts reasoner (975B total, 41B active) with controllable reasoning effort, fine-tunable on Tinker.

Open weights64K ctx · 64K outReasoningToolsVision
↓ 608K♥ 1.8K
15 Jul 2026

Bonsai 27B

PrismML

The first 27B-class model to run on a phone, based on Qwen3.6 27B: a multimodal flagship shipping in ternary (5.9 GB) and 1-bit (3.9 GB) forms with speculative decoding.

Open weights256K ctxReasoningToolsVision
Apache-2.0
14 Jul 2026

GPT-5.6 Luna

OpenAI

GPT-5.6 model optimized for cost-sensitive workloads.

API1.05M ctx · 128K outReasoningToolsVisionPDF
$0.2 / $1.2 per 1M
9 Jul 2026

GPT-5.6 Sol

OpenAI

Frontier model for complex professional work.

API1.05M ctx · 128K outReasoningToolsVisionPDF
$4 / $20 per 1M
9 Jul 2026

GPT-5.6 Terra

OpenAI

GPT-5.6 model that balances intelligence and cost.

API1.05M ctx · 128K outReasoningToolsVisionPDF
$2 / $12 per 1M
9 Jul 2026

Grok 4.5

xAI

xAI's Grok 4.5 flagship for chat, coding, and agentic tool use, with lower hallucination risk.

API500K ctx · 500K outReasoningToolsVisionPDF
$2 / $6 per 1M
8 Jul 2026

SWE-1.7

Cognition

The most capable model Cognition has trained so far. It reaches frontier-level intelligence at a much lower cost, advancing the cost-performance Pareto curve.

APITools
8 Jul 2026

Claude Sonnet 5

Anthropic

Claude Sonnet 5 is built to be the most agentic Sonnet model yet. It can make plans, use tools like browsers and terminals, and run autonomously at a level that, just a few months ago, required larger and more expensive models.

API1M ctx · 128K outReasoningToolsVisionPDF
$2 / $10 per 1M
29 Jun 2026

GLM-5.2

Zhipu AI

We're introducing GLM-5.2, our latest flagship model for long-horizon tasks.

Open weights1M ctx · 128K outReasoningTools

DeepSeek V4 Flash

DeepSeek

DeepSeek-V4-Flash-0731 is the official release of DeepSeek-V4-Flash, superseding the preview version, with substantially enhanced agentic capabilities.

Open weights1M ctx · 384K outReasoningTools
↓ 4M♥ 4K
24 Apr 2026

DeepSeek V4 Pro

DeepSeek

DeepSeek-V4-Pro is a 1.6T-parameter Mixture-of-Experts model with 49B activated, supporting a one-million-token context. The 0813 revision is its GA release.

Open weights1M ctx · 384K out1.6T total, 49B active (MoE)ReasoningTools

GPT-5.5

OpenAI

A new class of intelligence for coding and professional work.

API1.05M ctx · 128K outReasoningToolsVisionPDF
$5 / $30 per 1M
23 Apr 2026

Kimi K2.6

Moonshot AI

Kimi K2.6 is an open-source, native multimodal agentic model that advances practical capabilities in long-horizon coding, coding-driven design, proactive autonomous execution, and swarm-based task orchestration.

Open weights256K ctx · 256K out1TReasoningToolsVision
↓ 415K♥ 1.6K
21 Apr 2026

Beam

Reflection AI

Beam is a sparse Mixture-of-Experts model with 501 billion total parameters, 23 billion active, built for coding, reasoning, and agentic workloads.

API1M ctx

Clairic

Cloudflare

A System One extraction model for extracting, computing, estimating, and forecasting structured values from unstructured inputs. Available in preview on Cloudflare Workers AI.

API

Clef-flash

Cloudflare

A smaller decision model based on Qwen3.5-9B that scores schema-constrained options. The October update reduces hosted pricing and context size without changing the released weights.

Open weights9B

Clef-omni

Cloudflare

A multimodal decision model built on Qwen3-Omni-30B-A3B-Instruct. It scores schema-constrained options from text, images, audio, and video without generating free-form output.

Open weights30B total / 3B active (MoE)

ColGrep-Agent-2B

LightOn AI

A local code-search agent that searches repositories with ColGrep, reads relevant code, and returns file paths and line ranges to a coding agent. Post-trained from MiniCPM5-2B.

Open weights2B

d1-2B

Liquid AI

d1-2B is a decision model built on LFM3-VL-2B. It reads text and images and answers multiple questions in a single forward pass, each with a calibrated probability.

Open weights1.9B

d1-Omni-2B

Liquid AI

d1-Omni-2B is a decision model built on LFM3-Omni-2B. It reads text and images, listens to audio, and answers multiple questions in a single forward pass, each with a calibrated probability.

Open weights2.2B

DaisyRose-27B

RadixArk

RadixArk’s first model, built on Qwen3.8-27B and trained with reinforcement learning using Miles on long-horizon agent tasks.

Open weights27B dense

Gemini 3.5 Flash-Cyber

Google

Fine-tuned for finding and fixing cybersecurity vulnerabilities at a lower price per token than larger models.

API

Gemini 4 Neon

Google DeepMind

Bringing frontier-level intelligence with fast end-to-end task completion across complex software engineering, knowledge work, and science workflows.

API

Kimi K3.1 Pro

Moonshot AI

Kimi K3.1 Pro is a post-trained version of Kimi K3 with improvements in coding, 3D modeling, financial research, and agentic tasks. Available through Kimi's products and API, with full model weights scheduled for release by November 4, 2026.

API

Kimi K3.1 Pro Highspeed

Moonshot AI

Kimi K3.1 Pro Highspeed is the faster variant of Kimi K3.1 Pro, which Moonshot AI reports delivers 2–3 times the speed for more responsive interactions. Available through Kimi's products and API.

API

LFM2.5-24B-A2B

Liquid AI

LFM2.5-24B-A2B is a reasoning model with 24 billion total parameters and 2 billion active parameters, built for fast, private on-device agents. It combines gated short convolutions and attention with a mixture-of-experts architecture.

Open weights24B total, 2B active

LFM3-1.7B

Liquid AI

LFM3-1.7B is a text model for local agents and specialized applications.

Open weights1.7B

LFM3-Audio-2B

Liquid AI

LFM3-Audio-2B is a compact audio model that runs locally on laptops and edge devices. It takes text or audio as input and responds with text, speech, or both.

Open weights2B

LFM3-Omni-2B

Liquid AI

LFM3-Omni-2B brings vision, audio, and language together in a single on-device model that can see, listen, and talk back.

Open weights2.2B

LFM3-VL-2B

Liquid AI

LFM3-VL-2B is a vision-language model for video understanding, visual reasoning, and spatial and embodied reasoning on-device.

Open weights1.9B

Muse Wonder Full

Meta

A frontier model for personal and professional agents, combining Active Memory and Active Focus for sustained coding and complex tasks. Available to Muse Code subscribers.

API

Ornith 1.5 122B

Ornith

Ornith 1.5 122B is a mixture-of-experts reasoning model for coding and agentic workflows, with support for tool calling. Its 122 billion parameter weights are available under the MIT license.

Open weights122B

OUI-1

Thesys

An open-weight diffusion model fine-tuned from DiffusionGemma to generate user interfaces in OpenUI Lang on consumer GPUs.

Open weights26B total / 4B active

Recursive Flint

Recursive

A mixture-of-experts model built for fast agentic coding on local hardware, with 40 billion total parameters and 2.5 billion active per token. Built from scratch with Recursive’s Eureka Machine automated research system.

Open weights40B total, 2.5B active

SWE-2

Cognition

A coding model post-trained from Kimi K3 with reinforcement learning across reasoning effort levels. Available in Devin Desktop and CLI, with rollout to Devin Web and Fusion.

API2.8T

Ternary Bonsai 2 2B

PrismML

A ternary 2B model based on Qwen3.5, built for local agents with reasoning, visual understanding, and tool use.

Open weights2B

Ternary Bonsai 2 4B

PrismML

A ternary 4B model based on Qwen3.5, built for local agents with reasoning, visual understanding, and tool use.

Open weights4B

Ternary Bonsai 2 9B

PrismML

A ternary 9B model based on Qwen3.5, built for local agents with reasoning, visual understanding, and tool use.

Open weights9B

Ultramarine 1 Flash

Paper Instruments

An agentic model for professional knowledge work, post-trained from GLM-5.3 Flash with reinforcement learning on non-verifiable tasks. The announcement presents research results without a public model download or API release.

API

Keep up with the tools

An occasional email when notable AI dev tools and models land in the directory. No spam, unsubscribe anytime.