DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///DROPGoogle launches Gemini 3.8 Flash TTS and Flash-Lite TTS///DROPGemini Omni brings 1080p generation to Google Vids///BREAKClaude Opus 5.5 is here: 1M context, $4 input / $20 output///BREAKOpenAI adds GPT-6 Sol and Luna to the GPT-6 family///PAPEROpus 5.5's system card documents gains—and evaluation limits///BREAKGrok 4.7 launches at $2 input / $6 output per million tokens///DROPGrok Voice Transcribe 2.0 targets difficult real-world audio///TRENDGoogle steers new Gemini API projects to 3.8 Flash or 3.5 Flash-Lite///
DARC.DEV
Live from the lab/Independent AI intelligence
Model index · 2026-11-08 ↗

AI moves fast. Stay on frequency.

Signalin theNoise.

The personal website and independent AI developer lab of Dariusz Czechowicz. Software experiments. Practical guides. Source-linked research and useful tools.
Built and tested, not just imagined.

Neural signal

Live render
… FPS
Move to bend · drag to orbit · click a model
Inspecting / sourced

GPT-6.1 Terra

openai · 1.05M context

—DC Power
DC Score v3DARC.DEV · four source familiesChecked 2026-11-08
DC Power / capability · source & conditions

DC Power combines independent source families with fixed weights for Coding, Reasoning and Agents. Incomplete panels have no Power score. DC Eco estimates token cost. Inspect inputs, coverage and harness conditions in the model index.

DC multi-source weights and evidence ↗
Catalogue update delayed / 0011Full model index
Inspecting GPT-6.1 Terra: DC Power / capability. 35% Coding, 35% Reasoning, 30% Agents. Complete independent-source panel required..

Initializing neural field12 models · cited data

Primary-source reporting✦437 models in the catalogue✦No invented benchmark scores✦Working calculators & worksheets✦Sources on every briefing✦Read The Signal✦Primary-source reporting✦437 models in the catalogue✦No invented benchmark scores✦Working calculators & worksheets✦Sources on every briefing✦Read The Signal✦Primary-source reporting✦437 models in the catalogue✦No invented benchmark scores✦Working calculators & worksheets✦Sources on every briefing✦Read The Signal✦
[01 — Signal feed · source-linked intelligence]

What changed. Why it matters.

Current releases and platform changes, with the original source and a practical next step. Read the announcement, understand the constraint, decide what deserves a test.

Open the newsroom
Source-checked Dated editorial snapshot · official sources · no invented resultsSources & corrections
DROPGoogle ↗

Google launches Gemini 3.8 Flash TTS and Flash-Lite TTS

Google's 23 September release adds voice design, consent-verified voice replication and two-speaker direction. Flash TTS targets creative fidelity; Flash-Lite targets high-volume generation. Both are rolling out through Gemini API and AI Studio.

THE READ → Separate availability from the roadmap: voice remixing is described as coming soon, not shipping today.

DROPGoogle ↗

Gemini Omni brings 1080p generation to Google Vids

Google says anyone with a Google Account can create 1080p video in Vids with Gemini Omni. The update adds scene extension and clip-duration controls. Flash-Lite TTS narration is still marked coming soon in this announcement.

THE READ → A useful creator update, but a product rollout is not evidence that one video model wins every shot.

BREAKAnthropic ↗

Claude Opus 5.5 is here: 1M context, $4 input / $20 output

Released 22 September, Opus 5.5 replaces Opus 5 as the newest Opus. Standard token prices fall 20%; cache reads drop to $0.20/M. Anthropic reports 40% lower typical workload cost. Thinking is always on; Sonnet 5.5 and Haiku 5.5 are not yet released.

THE READ → The migration is more than a model-name edit: review always-on thinking, preserved context and safeguard fallback behavior.

BREAKOpenAI ↗

OpenAI adds GPT-6 Sol and Luna to the GPT-6 family

The 22 September API changelog lists GPT-6 Sol and GPT-6 Luna. Standard prices for prompts up to 272K input tokens are $2/$10 and $0.10/$0.50 per million input/output tokens respectively. Both accept text and images and return text.

THE READ → Astra, Sol and Luna are different tiers. Compare completed-task quality and the long-context surcharge, not just the headline rate.

PAPERAnthropic ↗

Opus 5.5's system card documents gains—and evaluation limits

The new system card covers agentic coding, computer use, alignment evaluations and deployment safeguards. Anthropic also reports weaknesses, including malicious instructions pasted into user prompts. Reported scores are not guarantees of safety in your application.

THE READ → Read the caveats alongside the headline benchmarks. A more robust model does not replace tool isolation and approval gates.

[03 — 29 tool profiles · documentation-reviewed]

Know the tool. Know the limits.

A useful tool profile is more than a logo and a rating. Get the use cases, a first workflow, cost drivers, privacy boundaries and reasons to choose something else.

Open tool radar

04 / DC Model Index

The next model.
An informed choice.

Latest releases from leading model providers, transparent DC scores and comparable token prices. Inspect source coverage, compare individual tests and estimate the cost of your workload.

Update delayed · last successful import · 437 models · Sources & methodology ↗

openai · Released 2026-10-26

GPT-6.1 Terra
—

DC Power

46.5% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

94.3

50% · partial

Reasoning

88.7

40% · partial

Agents

48.6

50% · partial

1.05M context · $1 / $5

USD per 1M input / output tokens

GPT-6.1 Terra is the mid-tier model in OpenAI's GPT-6.1 series, priced below GPT-6.1 Sol. It is suited for everyday agentic, coding, and writing work, is particularly strong at understanding screenshots...

Evidence & available offers →

anthropic · Released 2026-10-07

Claude Haiku 5.5
—

DC Power

67.5% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

91.4

50% · partial

Reasoning

74.2

Full panel

Agents

40.0

50% · partial

1M context · $0.1 / $0.5

USD per 1M input / output tokens

Claude Haiku 5.5 is Anthropic's small, fast model for high-volume, cost-sensitive work such as summarization, subagents, and browser use. It succeeds Claude Haiku 4.5 with stronger coding, computer use, and...

Evidence & available offers →

moonshotai · Released 2026-10-23

Kimi K3.1 Pro
—

DC Power

57% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

81.8

50% · partial

Reasoning

84.0

70% · partial

Agents

39.1

50% · partial

1.05M context · $2 / $10

USD per 1M input / output tokens

Kimi K3.1 Pro is Moonshot AI's flagship open-weight multimodal reasoning model and a post-trained successor to Kimi K3, with a 1M-token context window. It is built for long-horizon coding, 3D...

Evidence & available offers →

qwen · Released 2026-09-23

Qwen 3.8 Max Prime
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

1M context · $4 / $12

USD per 1M input / output tokens

Qwen3.8 Max Prime is a higher-throughput variant of Qwen3.8 Max from Alibaba's Qwen team, served as a separate SKU at a higher price point. It accepts text, image, and video...

Evidence & available offers →

google · Released 2026-11-05

Gemini Omni 1.2 Flash
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

131K context · $1.5 / $9

USD per 1M input / output tokens

Gemini Omni 1.2 Flash is Google's native multimodal video generation and editing model, the successor to Gemini Omni 1.1 Flash. It generates clips up to 15 seconds long with native...

Evidence & available offers →

deepseek · Released 2026-10-17

DeepSeek V4.1 Flash 1017
—

DC Power

67.5% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

96.0

50% · partial

Reasoning

80.7

Full panel

Agents

35.7

50% · partial

1.05M context · $0.006 / $0.5

USD per 1M input / output tokens

DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek V4.1 Flash, served by DeepSeek under the `deepseek-flash` API model name. V4.1 Flash is a sparse mixture-of-experts model...

Evidence & available offers →

x-ai · Released 2026-09-21

Grok 4.7
—

DC Power

82.5% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

84.5

50% · partial

Reasoning

70.7

Full panel

Agents

41.8

Full panel

500K context · $2 / $6

USD per 1M input / output tokens

Grok 4.7 is SpaceXAI's flagship model for coding, agentic tasks, and knowledge work, succeeding Grok 4.6. It is particularly strong at long-running software engineering tasks, verifying its own work, and...

Evidence & available offers →

mistralai · Released 2026-10-06

Mistral Large 4
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

1.05M context · $0.7 / $2.1

USD per 1M input / output tokens

Mistral Large 4 is a frontier multimodal (text and image input) model from Mistral AI built for reasoning, coding, and agentic workloads. It offers a 1M-token context window with up...

Evidence & available offers →

meta · Released 2026-10-15

Muse Spark Open
—

DC Power

57% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

73.6

50% · partial

Reasoning

66.8

70% · partial

Agents

11.7

50% · partial

1.05M context · $1.25 / $4.25

USD per 1M input / output tokens

Muse Spark Open is a coding-optimized open-weight model. Trained specifically for agentic coding workflows, it delivers higher first-attempt accuracy, reliable tool calling, and deep codebase understanding across the repos developers...

Evidence & available offers →

z-ai · Released 2026-11-02

GLM 5.3 FlashX Legal
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

1.05M context · $0.37 / $1.25

USD per 1M input / output tokens

GLM-5.3-FlashX-Legal (GLM-Legal) is Z.ai's legal-domain variant of GLM-5.3-FlashX, a native multimodal model that accepts text, image, and video input and produces text. It adds legal-domain mid-training and post-training with reinforcement...

Evidence & available offers →

liquid · Released 2026-11-06

LFM3-1.7B
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

262K context · $0.08 / $0.2

USD per 1M input / output tokens

A compact 1.7B-parameter text model with a 256K-token context, post-trained for agents, tool calling, and structured outputs. It reasons before it answers. Prompts and outputs may be retained and used...

Evidence & available offers →

tencent · Released 2026-11-04

Hy4
—

DC Power

0% fixed panel · Power withheld
—

DC Eco

—

DC Overall

Coding

—

0% · partial

Reasoning

—

0% · partial

Agents

—

0% · partial

1.05M context · $0.417 / $1.2505

USD per 1M input / output tokens

The official release of Hy4 delivers comprehensive and substantial improvements over Hy4 Preview, with particularly strong gains in agentic capabilities and performance across a wide range of long-horizon, real-world productivity...

Evidence & available offers →
Explore the live model index →
[06 — Watch & inspect · original sources]

See the mechanics. Question the demo.

Official demonstrations and developer walkthroughs, paired with a viewing brief. Know what to look for, what is actually available and what a polished example cannot prove.

Watch & read

An editorial principle

The model will change.
The ability to ask, build and verify
is the part worth keeping.

DARC.DEV is my independent developer lab, where AI research becomes practical skills, software experiments and working tools. The focus is on evidence, design and testing.

The Signal · web editions available

Keep the signal.
Lose the noise.

A compact briefing that connects a release to a decision: what changed, what the source says and what to try next. Read every current edition on the site.

Email interest · no automatic send

Your registration is stored in this application’s database. Email delivery is not configured; no message is sent by this form.