ModelBrief · Models · Topic Page
Last updated Nov 20, 2026
22 topics · 17 vendors

All Models

An AI model comparison across all 22 ModelBrief topics — 16 LLMs, 3 image models and 3 video models from 17 vendors, newest first. The table below is an LLM pricing comparison with context windows and open-weight status, sourced from each topic page; open a card below for the full review.

01Comparison table

One row per model — price, context and open-weight status side by side. Open a model name for the full review and pricing breakdown.

ModelVendorRelease dateContext or scaleInput · Output priceOpen weights
Gemini 4 NeonGoogleNov 18, 20261M tokens$3 / $15 per 1M tokensNo
GPT-6.2 SolOpenAINov 17, 20261M tokens$2 / $10 per 1M tokensNo
GPT-6.2 AstraOpenAINov 17, 20261M tokens$8 / $40 per 1M tokensNo
MAI Code 2MicrosoftNov 17, 20261.05M tokens$0.50 / $2.50 per 1M tokens (promo to May 17, 2027)No
Gemini 4 HydrogenGoogleNov 12, 20261M tokens$1.50 / $7.50 per 1M tokensNo
Seedance 2.5 FlashByteDanceNov 12, 2026 (API Nov 15)Video · up to 30s · 720p¥0.99 / second (720p)No
Edge0-145BEdge0Nov 12, 2026145B total / 10.12B active · 256KNo API (local only)Weights public (custom license)
Qwen4-MaxAlibabaNov 11, 2026 (GA Nov 16)1M tokens$1.50 / $4.50 per 1M tokensNo
ERNIE 5.5 Flash PreviewBaiduNov 11, 20261M tokens¥1–2 / ¥4–8 per 1M tokens (Qianfan peak/off-peak)No (weights promised by Nov 25)
Gemini 4 CarbonGoogleNov 10, 20261M tokens$2 / $10 until Nov 30
$4 / $20 from Dec 1 (per 1M tokens)
No
MiniMax H3-ImageMiniMaxNov 10, 2026Image · 33.1B · native 4K$0.06879 / image (4K T2I example)Weights public (custom license)
Nex-N2.6Nex-AGINov 10, 2026262K (OpenRouter) / 1M (release config)$0.12 / $0.40 per 1M tokens (20% off)Yes (Apache 2.0)
Seedream 5.5ByteDanceNov 9, 2026Image · up to 2K¥0.13–¥0.23 / image (output)No
Claude Fable 5.2AnthropicNov 9, 20261M tokens$8 / $40 per 1M tokensNo
Muse EmberMetaNov 9, 20261M tokens$0.10 / $0.50 per 1M tokens
(≤200K input; 3× above 200K)
No
Inkling 2 SmallThinking Machines LabNov 5, 2026 (weights Nov 12)1M tokens (arch. cap; Tinker 64K/256K)$0.30 / $1.00 per 1M tokensYes (Apache 2.0)
Gemini Omni 1.2 FlashGoogleNov 5, 2026Video · 3–15 s~$0.07 / s (720p)No
Kimi K3.1 ProMoonshot AINov 5, 2026 (weights; API Oct 23)1M tokens$2 / $10 per 1M tokensYes (custom license)
Tencent Hunyuan Hy4TencentNov 4, 20261M tokens¥6 / ¥18 per 1M tokens
(list price; 50% off to Dec 31)
No (preview weights only)
Kling 4.0 FlashKuaishouNov 4, 2026 (GA; announced Sep 28)Video · up to 20s · 720p$0.094 / second (720p)No
GPT Image 2.6OpenAINov 4, 2026Image · up to 4K (tested)$0.0247–$0.2107 / image
(1024×1024, low–max)
No
LongCat 2.5MeituanNov 2, 20261M tokens$0.30 / $1.20 per 1M tokens
(promo, no end date)
Yes (MIT)

Pricing basis: LLMs are priced per 1M tokens (input · output); image models per image (MiniMax H3-Image and GPT Image 2.6 actually bill per token — the table shows official per-image examples); video per second (Gemini Omni 1.2 Flash actually bills per token — the table shows the official per-second conversion). RMB and USD prices are separate vendor price lists, not directly comparable; the table shows the primary channel only — see each topic page for other channels. Prices change often; check the official pages. "—" means the topic page does not state it. Nex-N2.6 context has two figures: 262,144 tokens on the OpenRouter endpoint, 1,048,576 in the release config.

02All topics

Newest first, in release order — open a card to read the full topic page.

USA · LLM · GA · Nov 18, 2026Gemini 4 NeonReleased Nov 18, 2026 · Google's mid-tier Gemini 4 model, AA Index 61 (3rd), #1 on the AA Coding Agent Index, $3/$15 per 1M tokens, 1M context. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. USA · LLM · GA · Nov 17, 2026GPT-6.2 SolReleased Nov 17, 2026 · OpenAI's everyday model, AA Index 57, $2/$10 per 1M tokens, 1M context. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. USA · LLM · GA · Nov 17, 2026GPT-6.2 AstraReleased Nov 17, 2026 · OpenAI's flagship reasoning model, AA Index 62 in the top 3, $8/$40 per 1M tokens, 1M context. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. USA · LLM · GA · Nov 17, 2026MAI Code 2Released Nov 17, 2026 · Microsoft's first in-house coding flagship, live in GitHub Copilot on day one; Microsoft claims 98% of GPT-6.1 Sol's performance at 40% of the cost (50% promo price). The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. USA · LLM · GA · Nov 12, 2026Gemini 4 HydrogenReleased Nov 12, 2026 · The fast Gemini 4 tier, succeeding 3.8 Flash, $1.50/$7.50 with a free tier. The topic page covers official specs, pricing and 9 in-depth reviews. China · Video · GA · Nov 12, 2026Seedance 2.5 FlashReleased Nov 12, 2026 · ByteDance's cheap Seedance 2.5 tier, ¥0.99/s at 720p, ~34% below the flagship, API live Nov 15. The topic page covers official specs, pricing, comparisons and 7 in-depth reviews. LLM · Open weights · Released Nov 12, 2026Edge0-145BReleased Nov 12, 2026 · Edge0's first 145B on-device model, a latent MoE on a Qwen3.8-Flash-Next base, 256K context, weights public on HF under a custom community license (not open source). The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. China · LLM · GA · Nov 16, 2026Qwen4-MaxReleased Nov 11, 2026 · Alibaba's first Qwen4 flagship, generally available Nov 16 with pay-as-you-go API at $1.50/$4.50; preview ends Nov 25. The topic page covers official specs, Token Plan pricing and 18 in-depth reviews. USA · LLM · GA · Nov 10, 2026Gemini 4 CarbonReleased Nov 10, 2026 · Google's strongest model, built for long-horizon agents and cybersecurity. The topic page covers official specs, API pricing and 8 in-depth reviews. China · Image · GA · Nov 10, 2026MiniMax H3-ImageReleased Nov 10, 2026 · MiniMax's first image model — native 4K at 16.8MP, 6-step sampling, public weights. The topic page covers official specs, pricing, license and 8 in-depth reviews. China · LLM · Open weights · Nov 10, 2026Nex-N2.6Released Nov 10, 2026 · Nex-AGI's computer-use agent, post-trained on GLM-5.3-Flash, OSWorld 83.0, Apache 2.0 open FP8 weights. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. China · LLM · Preview · Nov 11, 2026ERNIE 5.5 Flash PreviewReleased Nov 11, 2026 · Baidu's 280B/14B MoE, 1M context, Qianfan API live, free tier on OpenRouter. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. China · Image · GA · Nov 9, 2026Seedream 5.5Released Nov 9, 2026 · ByteDance's image model — 100-round editing and native transparent PNG, AA T2I 1135, editing 1159. The topic page covers official specs, pricing, API status and 7 in-depth reviews. USA · LLM · GA · Nov 9, 2026Claude Fable 5.2Released Nov 9, 2026 · Anthropic's flagship reasoning model, AA Index 65, No. 1 on the board, $8/$40 per 1M tokens. The topic page covers official specs, pricing, an effort-tier guide and 7 in-depth reviews. USA · LLM · GA · Nov 9, 2026Muse EmberReleased Nov 9, 2026 · Meta's smallest, cheapest Muse model, AA Index 36 matching Claude Haiku 5.5 at a third of the task cost. The topic page covers official specs, pricing, benchmark analysis and 6 in-depth reviews. USA · LLM · Open weights · Nov 12, 2026Inkling 2 SmallReleased Nov 5, 2026 · Thinking Machines Lab's 594B open MoE, Apache 2.0 weights out Nov 12. The topic page covers official specs, Tinker pricing and 6 in-depth reviews. USA · Video · GA · Nov 5, 2026Gemini Omni 1.2 FlashReleased Nov 5, 2026 · Google's video model — native synced audio, editable 3–15s video, 720p ~$0.07/s. The topic page covers official specs, pricing, Flow access and 6 in-depth reviews. China · LLM · Open weights · Nov 5, 2026Kimi K3.1 ProReleased Nov 5, 2026 · Moonshot's 2.8T open-weight MoE, AA Index 46 (v5.0), LMArena 1488, $2/$10. The topic page covers official specs, pricing, benchmark analysis and 7 in-depth reviews. China · LLM · GA · Nov 4, 2026Tencent Hunyuan Hy4Released Nov 4, 2026 · Tencent Hunyuan flagship MoE, 770B total params, AA Index 35, TokenHub ¥6/¥18 at 50% off through Dec 31. The topic page covers official specs, pricing and 12 in-depth reviews. China · Video · GA · Nov 4, 2026Kling 4.0 FlashGenerally available Nov 4, 2026 · Kuaishou's cheap, fast Kling 4.0 tier, 720p at $0.094/s, ~40% below the full model, 20s cap. The topic page covers official specs, pricing, comparisons and 6 in-depth reviews. USA · Image · GA · Nov 4, 2026GPT Image 2.6Released Nov 4, 2026 · OpenAI's image model — Flare and Sunburst variants, new GIF and SVG export. The topic page covers official specs, pricing, API access and 6 in-depth reviews. China · LLM · GA · Nov 2, 2026LongCat 2.5Released Nov 2, 2026 · Meituan's 1.6T native multimodal MoE, $0.30/$1.20 promo pricing, MIT weights out Nov 4. The topic page covers official specs, pricing and 6 in-depth reviews.