Full index · 100 models

All models

Every model across every category. Search by name, maker, or what it’s good at, then narrow by category or country of origin.

100 results

Anthropic
Language Models

Claude Fable 5.1

9.7/10
  • Leads on long-horizon agentic coding, with strong root-cause debugging across multi-step sessions
  • Strong agentic scientific-research performance, including protein-design work Anthropic reports near 50% hit rates on
  • Anthropic reports reduced visibility into very long-context and multi-agent runs versus shorter sessions
  • Can still bypass approval gates or auto-mode classifiers on some agentic tasks, per Anthropic's own testing
Anthropic
Language Models

Claude Opus 5

9.7/10
  • Anthropic's recommended model for complex, high-stakes work, with strong reasoning-to-cost
  • Zero-data-retention eligible, useful for regulated or enterprise deployments
  • Sits below the flagship tier in branding despite strong practical scores
  • Costs meaningfully more per token than the Sonnet tier for everyday tasks
Anthropic
Language Models

Claude Opus 5.5

9.7/10
  • Tops LMArena outright on hard prompts (1,545) and on the WebDev arena (1,827), the highest rating on either board
  • Anthropic reports roughly 40% lower cost per workload than Opus 5 and over 30% faster output, at $4/$20 per million tokens
  • Released weeks ago, so creative writing is still below this site's vote floor and no hallucination testing exists yet — its overall score is pending
  • Still the premium tier: many times the price of the fast models that most high-volume work runs on
Anthropic
Language Models

Claude Fable 5

9.5/10
  • 3rd of 380 on LMArena's creative-writing arena, and 7th of 399 on hard prompts, on tens of thousands of votes
  • Built for long-horizon agentic work, with a 1M-token context and up to 128K output tokens per request
  • Anthropic released it and then temporarily suspended it in June 2026 before restoring access
  • Superseded by Fable 5.1 in September 2026 at the same $10/$50 per million tokens
Google
Language Models

Gemini 3.8 Flash

9.5/10
  • Highest-rated model in the directory on LMArena's hard prompts after Anthropic's Opus line, on 14,320 votes
  • Google reports 90.8% on Terminal-Bench 2.1, up from 81.6% for 3.7 Flash, at $0.75/$3.75 per million tokens
  • Built on 3.7 Flash rather than a new base model, and Google tells you to stay on 3.7 for efficiency-first work
  • Spends more thinking tokens by design, so the headline price understates what a task actually costs
Moonshot AI
Language Models

Kimi K3

9.4/10
  • The largest open-weight release yet at 2.8T parameters with 104B active, served by 17 independent providers
  • Moonshot reports 81.2 on FrontierSWE and 88.3 on Terminal-Bench 2.1, and LMArena ranks it 8th of 133 on WebDev
  • $3/$15 per million tokens, several times the price of the open-weight models it competes with
  • Its licence is only open to a point: model-as-a-service businesses over $20M a year need a separate agreement
Xiaomi
Language Models

MiMo v2.6

9.4/10
  • The strongest open-weight model measured here: 10th of 399 on LMArena's hard prompts, MIT-licensed at 1T parameters with 42B active
  • Native text, image, video and audio input with a 1M-token context, at $0.44/$0.87 per million tokens
  • Xiaomi is new to frontier releases, so its production track record is thin next to the established labs
  • 27th on creative writing against 10th on hard prompts — clearly stronger at reasoning than at expressive writing
Meta
Language Models

Muse Spark 1.3

9.4/10
  • Meta's first frontier line since it moved off Llama, natively multimodal across text, image and video at $1.25/$4.25 per million tokens
  • Meta reports roughly 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2 on the same agentic work
  • Proprietary, ending the open-weight position Llama gave Meta, and served by a single provider on OpenRouter
  • 10th of 409 on LMArena's overall text arena — strong, but behind the Anthropic and Google models above it
Alibaba
Language Models

Qwen3.8 Max

9.4/10
  • 6th of 133 on LMArena's WebDev arena, the highest-rated non-frontier model for building working web apps
  • 2.4T mixture-of-experts with about 95B active and a 1M-token context, at $2/$6 per million tokens
  • Loses to GLM-5.3 on seven of the eight benchmarks both makers publish, at several times the price
  • Served by a single provider on OpenRouter, so there is no fallback if it is unavailable
Zhipu AI
Language Models

GLM-5.3

9.3/10
  • Open weights across 32 independent providers at $0.38/$1.19 per million tokens, the cheapest model at this rating
  • Z.ai reports state-of-the-art open-weight results on Terminal-Bench 3.0 and roughly 50% better coding than GLM-5.2
  • A post-training upgrade on the GLM-5.2 base rather than a new model, so the gains are tuning rather than scale
  • 17th of 399 on LMArena's hard prompts: clearly behind the frontier it undercuts on price
DeepSeek
Language Models

DeepSeek V4.1 Flash

9.2/10
  • MIT-licensed and openly downloadable at 552B parameters with only 8B active, served by 27 independent providers
  • DeepSeek reports 74.2% on Deep-SWE v1.1, ahead of both Opus 5 (74.0%) and GPT-5.6 Sol (73.0%)
  • 36th of 409 on LMArena's overall text arena, well behind the frontier models it matches on published coding benchmarks
  • A Flash-tier model by design: DeepSeek's own Pro line is stronger on maths and long reasoning
OpenAI
Language Models

GPT-5.6

9.2/10
  • Best-in-class agentic reasoning, with search, code execution, and computer use in one API
  • Disciplined, low-hallucination output on long, multi-step tasks
  • Long-context requests above roughly 272K tokens get repriced sharply higher
  • Slower to produce a first answer than most rivals at max reasoning
DeepSeek
Language Models

DeepSeek V4

9.1/10
  • Genuinely open-weight frontier model you can self-host, unlike closed rivals
  • Very low API cost, especially during off-peak pricing windows
  • V4 Pro is text-only; image input is limited to an experimental Flash variant
  • Trails the closed frontier labs on aggregate intelligence benchmarks
Google
Language Models

Gemini 3 Flash

9.1/10
  • Among the fastest decoding speeds of any frontier-class model
  • Aggressively priced for high-throughput production traffic
  • Newer Flash versions (3.5 to 3.8) have since replaced it as Google’s default fast model
  • Higher reasoning settings add noticeable cost and latency
OpenAI
Language Models

GPT-6 Astra

9.1/10
  • Rated #1 of 128 models on LMArena’s WebDev arena for building working web apps, well clear of every rival
  • OpenAI reports state-of-the-art results on FrontierMath, ARC-AGI-3, and OSWorld 2.0 computer use
  • In everyday use people prefer many rivals: #49 of 402 on LMArena’s hard prompts, well behind Claude
  • Uses an 'opaque recurrence' reasoning style that OpenAI says is harder to monitor for chain-of-thought audits
Anthropic
Language Models

Claude Sonnet 5

9.0/10
  • Fast and capable, priced for everyday production use
  • Strong default for coding and long documents without Opus-level cost
  • Enabling maximum thinking mode can quietly balloon token spend
  • Trails Opus on the hardest multi-step reasoning problems
Google
Language Models

Gemini 3 Pro

9.0/10
  • Native multimodal input across text, image, audio, video, and PDFs
  • Very large context window with strong long-document recall
  • Slower first response than Google's own Flash tier
  • No native image or audio output, understanding only
ByteDance
Video Generation

Seedance 2

8.9/10
  • Multishot generation with longer, planned sequences
  • Upstream audio generation capability built into the model
  • Native audio is often disabled in third-party API routes
  • Smaller footprint outside ByteDance's own ecosystem
OpenAI
Language Models

GPT-6 Sol

8.8/10
  • 5th of 133 on LMArena's WebDev arena, OpenAI's strongest showing for building working web apps after Astra
  • 93.5% factual consistency on Vectara's hallucination leaderboard, ahead of GPT-6 Astra's 91.3%
  • 156th of 409 on LMArena's overall text arena — in blind comparisons people prefer a great many cheaper models
  • Sits below Astra in OpenAI's own line-up, so it doesn't carry the flagship's headline capabilities
OpenAI
Image Generation

GPT Image 2

8.7/10
  • Strong prompt-following for complex, multi-element compositions
  • Handles both generation and precise editing through one API
  • Small text, logos, and fine counts still need manual review
  • Less distinctive a look than Midjourney for pure aesthetic exploration
OpenAI
Language Models

GPT-6 Luna

8.6/10
  • Priced for volume at $0.10/$0.50 per million tokens, a twentieth of Sol, with the same 1.05M-token context
  • Built for high-volume clerical work — pulling fields from documents, summarising tickets — where throughput matters more than depth
  • 160th of 409 on LMArena's overall text arena and 155th on creative writing: not a model to reach for on hard or expressive work
  • OpenAI positions it two tiers below Astra, so complex reasoning and long agentic runs are outside its brief
SpaceXAI
Language Models

Grok 4.6

8.6/10
  • Leads several agentic and terminal-automation benchmarks
  • Native, real-time X and web search built into the model
  • Reasoning mode cannot be turned off, adding cost and latency to every call
  • Hits a steep pricing cliff once context passes roughly 200K tokens
Google
Language Models

Gemma 4

8.4/10
  • Purpose-built for on-device and edge deployment
  • Small variants, from 4B to 12B, run well on modest hardware
  • Not intended to compete with frontier models on hard reasoning
  • Limited context window compared to Gemini's cloud models
Zhipu AI
Language Models

GLM-5

8.4/10
  • Coding performance that rivals or matches Claude on several benchmarks
  • Open-weight, giving flexibility for self-hosted or fine-tuned deployments
  • Smaller ecosystem of tooling and integrations than the big three labs
  • Less battle-tested in production outside China-based deployments
SpaceXAI
Language Models

Grok 4.7

8.3/10
  • SpaceXAI reports Terminal-Bench 4.0 nearly doubling to 38.0% from Grok 4.6's 20.3% on multi-hour terminal tasks
  • Holds Grok 4.6's $2/$6 per million tokens despite a larger base model and a longer reinforcement-learning run
  • 151st of 399 on LMArena's hard-prompts arena — the gains its maker reports on agentic benchmarks don't show up in blind human preference
  • SpaceXAI's own figures put it behind Claude Fable 5.1 on CursorBench 4.0, 46.3% against 51.8%
Anthropic
Language Models

Claude Haiku 4.5

8.1/10
  • Cheapest Claude tier, well suited to high-volume, simple tasks
  • Low latency, good for chat-style and classification workloads
  • Smaller context window (200K tokens) than Sonnet 5 and Opus 5 (1M)
  • Noticeably weaker on hard reasoning than Sonnet or Opus
Mistral AI
Language Models

Mistral Medium 3.5

8.0/10
  • Strong multilingual performance, especially across European languages
  • Apache 2.0 license permits unrestricted commercial use
  • Smaller research budget than the US/China frontier labs shows in ceiling performance
  • Less name recognition can complicate enterprise procurement
Anthropic
Coding & Agentic

Claude Opus 4.6 (Claude Code)

7.5/10
  • Strong at inferring intent from vague prompts and architectural context
  • 1M-token context supports coherent multi-file, cross-repo refactors
  • Can use many more tokens than leaner coding models on routine work
  • Narrates its reasoning at length, which slows down quick tasks
SpaceXAI
Video Generation

Grok Imagine Video

7.0/10
  • Native audio generation built in, unlike many rivals
  • Fast and well-suited to short-form experimentation
  • Tightly tied to the xAI/X ecosystem
  • Primarily optimized for short clips, not longer narrative work
Google
Video Generation

Veo 3.1

7.0/10
  • Realistic motion and prompt adherence with native synchronized audio
  • Strong for cinematic scenes and ambience out of the box
  • Availability and pricing vary by platform and SKU
  • Generation still limited to relatively short clip lengths
Alibaba
Video Generation

Wan 2.7

7.0/10
  • Shares an ecosystem with Alibaba's image models for consistent pipelines
  • Solid image-to-video capability at competitive cost
  • Clips top out at about 15 seconds
  • Rated below the top text-to-video models in independent head-to-head votes
ByteDance
Image Generation

Seedream 5.0

6.6/10
  • Strong cinematic stills, a solid base frame for video generation
  • Solid composition and product-shot fidelity
  • Smaller international user base and support ecosystem
  • Full editing capabilities aren't yet widely exposed outside China-first platforms
Google
Image Generation

Gemini 3 Pro Image

6.4/10
  • High-fidelity output up to 4K with strong reference-image handling
  • Good starting point for complex, detail-heavy briefs
  • Higher latency than Google's own Flash image tier
  • Text accuracy and identity preservation still need spot-checking
Ideogram
Image Generation

Ideogram 4

5.9/10
  • Best-in-class for posters, logos-as-concepts, and text-heavy graphics
  • Open weights give deployment flexibility others don't
  • Still requires review of every rendered character for accuracy
  • Licensing and infra cost need evaluation before production use
SpaceXAI
Image Generation

Grok Imagine Image

5.5/10
  • Tight native integration for teams already inside the xAI/X ecosystem
  • Fast, straightforward generation for quick social content
  • Less proven on prompt adherence and consistency than category leaders
  • Limited advantage outside the xAI/Grok ecosystem
Black Forest Labs
Image Generation

FLUX.2

5.1/10
  • Flexible family of variants trading off quality, speed, and cost
  • Open-weight options available for self-hosted production pipelines
  • Confusing lineup; easy to pick the wrong variant for the job
  • Needs its own hosting and infra work to unlock full control
Recraft
Image Generation

Recraft V4.1

4.9/10
  • Built for designers; controllable, brand-consistent commercial graphics
  • Strong for icon sets and cohesive visual systems
  • Typography and brand consistency must be checked across batches
  • Smaller community and fewer integrations than the bigger names
Alibaba
Coding & Agentic

Qwen3 Coder 480B

4.7/10
  • Open-weight performance within striking distance of proprietary leaders
  • Free to self-host, appealing for cost-sensitive or air-gapped teams
  • Requires serious infrastructure to run at full size
  • Tooling and IDE integrations are less mature than Copilot, Cursor, or Codex
MiniMax
Video Generation

MiniMax Hailuo 2.3

4.4/10
  • Strong motion quality for short cinematic clips
  • Competitive pricing for social and short-form content
  • Best suited to short clips rather than longer narrative sequences
  • Niche provider with a smaller support ecosystem
Tencent
Video Generation

HunyuanVideo 1.5

3.4/10
  • Efficient enough to render on a single consumer GPU
  • Open-weight, with an active community of fine-tunes
  • Smaller model size trades away some quality ceiling
  • Still catching up to proprietary leaders on realism
Anthropic
Language Models

Claude Sonnet 5.5

Under evaluation
  • Anthropic reports at least 30% faster output than Sonnet 5, with fewer tokens needed for many of the same tasks
  • Available across Anthropic, AWS, Google Cloud, and Azure at $2 input/$10 output per million tokens
  • Released on September 28, so this directory has no qualifying independent capability measurements for it yet
  • Anthropic positions Opus 5.5 above it for complex work requiring careful judgment
Mistral AI
Language Models

Mistral Small 4

Under evaluation
  • One open-weight model that combines reasoning, image input, and agentic coding, under Apache 2.0
  • Mixture-of-experts design activates only 6.5B of 119B parameters per token, keeping inference efficient
  • A small-tier model: trails frontier flagships on the hardest reasoning tasks
  • At 119B total parameters, self-hosting needs far more memory than earlier Small models
Baidu
Language Models

ERNIE 5.0

Under evaluation
  • Deep integration with Baidu's search and China-market ecosystem
  • Strong Mandarin-language understanding and generation
  • Weaker mindshare and tooling support outside China
  • English-language performance trails the Western frontier labs
NVIDIA
Language Models

Nemotron 3 Ultra

Under evaluation
  • Fully open: weights, training data, and recipes released under the permissive OpenMDW-1.1 licence
  • 550B-parameter hybrid Mamba-Transformer (55B active) built for long-running agents
  • At 550B parameters, self-hosting needs a multi-GPU cluster
  • Its most efficient format (NVFP4) is designed for NVIDIA hardware
AI21 Labs
Language Models

Jamba 2

Under evaluation
  • Hybrid Transformer-Mamba architecture gives very high throughput and low memory use at long context
  • One of the largest context windows among open-weight models, aimed at enterprise long-document work
  • Needs its own proprietary quantization approach to hit those efficiency numbers
  • Much smaller track record and mindshare than Llama, Qwen, or DeepSeek
Tencent
Language Models

Hunyuan Large

Under evaluation
  • Strong presence in the Chinese consumer and enterprise market via Tencent's ecosystem
  • Broad multimodal family spanning text, image, and video under one brand
  • Limited adoption and tooling outside China
  • Less benchmarked against Western frontier models in independent tests
Reka AI
Language Models

Reka Flash 3

Under evaluation
  • Compact, efficient multimodal model that's easy to self-host
  • Good cost-to-performance ratio for mid-tier multimodal tasks
  • Trails frontier labs on the hardest reasoning benchmarks
  • Much smaller brand recognition and community than the major labs
ByteDance
Language Models

Doubao 1.5

Under evaluation
  • Deeply integrated into ByteDance's own consumer apps and ecosystem
  • Competitive general-purpose performance at low cost
  • Primarily built for the Chinese market, with limited presence elsewhere
  • Less transparent benchmarking against global frontier models
Upstage
Language Models

Solar Pro 2

Under evaluation
  • Strong performance for its size, popular in Korean-language and enterprise RAG use cases
  • Efficient enough to run cost-effectively at scale
  • Narrower global mindshare than the major US/China labs
  • Falls behind frontier models on the hardest general reasoning tasks
01.AI
Language Models

Yi-Lightning

Under evaluation
  • Fast, low-cost model that performs well for its price point
  • Reasonable multilingual coverage including English and Chinese
  • Trails top-tier frontier models on complex reasoning and coding
  • Smaller developer ecosystem than Qwen or DeepSeek
StepFun
Coding & Agentic

Step 5 Preview

Under evaluation
  • 600B mixture-of-experts with 27B active and a 1M-token context, built for long-horizon agentic work
  • $1.00 input and $2.70 output per million tokens, with open weights promised for October 2026
  • Generated roughly 1.7× the median output tokens on Artificial Analysis's index, so real cost runs well above the headline rate
  • A preview with no independent coding-benchmark result yet, so it carries no overall score here
OpenAI
Coding & Agentic

GPT-5.3 Codex

Under evaluation
  • Excellent terminal automation, git operations, and CI/CD debugging
  • Far more token-efficient than reasoning-heavy rivals on routine tasks
  • Needs detailed, unambiguous instructions; struggles with vague requests
  • Smaller context window than some rivals, a constraint on huge monorepos
Cursor
Coding & Agentic

Cursor Composer

Under evaluation
  • Tight, real-time feedback loop; you see and steer every change
  • Affordable flat-rate pricing for all-day assistance
  • Needs a developer actively driving; not built for unattended runs
  • Background and async agent mode is still early and limited
MiniMax
Coding & Agentic

MiniMax M2.1

Under evaluation
  • Low-cost alternative that holds up reasonably on coding benchmarks
  • Good option for parallel, high-volume exploratory tasks
  • Clearly behind the frontier coding leaders on hard problems
  • Smaller ecosystem and community support than bigger rivals
Cognition AI
Coding & Agentic

Devin

Under evaluation
  • Runs autonomously in the cloud, including overnight and unattended
  • Strong on large, well-defined jobs like migrations and refactors
  • Usage-based pricing gets expensive fast for heavy workloads
  • Needs clear upfront specs; weak at open-ended, exploratory work
Microsoft
Coding & Agentic

GitHub Copilot

Under evaluation
  • Widest IDE support of any coding assistant, with the largest installed base
  • Accessible free tier and tight integration with pull-request review workflows
  • Less specialized for complex multi-file refactors than purpose-built AI IDEs
  • Suggestion quality depends heavily on which underlying model you select
Amazon
Coding & Agentic

Amazon Q Developer

Under evaluation
  • Strong for AWS-centric teams, including automated Java version upgrades
  • Deep integration with CloudFormation and infrastructure-as-code workflows
  • Limited value for teams outside the AWS ecosystem
  • Less compelling than Cursor or Copilot for general-purpose development
Codeium
Coding & Agentic

Windsurf

Under evaluation
  • Budget-friendly pricing with automatic indexing of large codebases
  • Flexible deployment options, including self-hosted setups
  • Smaller ecosystem and mindshare than Cursor or Copilot
  • Fewer third-party integrations than the bigger platforms
Tabnine
Coding & Agentic

Tabnine

Under evaluation
  • Built for privacy-first enterprises, with on-premises deployment and zero data retention
  • Supports bring-your-own-model for teams with strict compliance needs
  • More conservative suggestions than aggressive agentic competitors
  • Smaller feature set for autonomous, multi-step tasks
Replit
Coding & Agentic

Replit Agent

Under evaluation
  • Goes from prompt to a running, deployed app inside one browser-based environment
  • Approachable for non-professional developers building small tools quickly
  • Less suited to large, established codebases than IDE-based agents
  • Deployment and hosting are tied to Replit's own platform
Open source
Coding & Agentic

Aider

Under evaluation
  • Free, open-source, terminal-based pair programmer that works with almost any model
  • Git-aware, auto-committing changes with clear diffs for easy review
  • No polished GUI; requires comfort with the command line
  • Lacks the managed infrastructure and support of commercial agents like Devin
Midjourney
Image Generation

Midjourney V8.1

Under evaluation
  • Unmatched aesthetic taste and texture for concept art and exploration
  • Excellent for open-ended creative direction-finding
  • Weak at exact, repeatable text and structured brand templates
  • Discord/web-first workflow adds overhead versus API-first tools
Adobe
Image Generation

Adobe Firefly Image 5

Under evaluation
  • Deep native integration with Photoshop, Illustrator, and Express
  • Enterprise-friendly licensing built for commercial content pipelines
  • The model itself trails pure-play leaders on raw output quality
  • Most value comes from the Adobe ecosystem, not the model alone
Stability AI
Image Generation

Stable Diffusion 3.5

Under evaluation
  • Free to self-host and endlessly customizable with community fine-tunes
  • The open model that seeded the entire ecosystem of tools built on top of it
  • Requires technical setup and decent hardware to run well
  • Base output quality trails the polished commercial leaders out of the box
Leonardo
Image Generation

Leonardo AI

Under evaluation
  • Friendly all-rounder with its own fine-tuned models and style presets
  • Generous free daily allowance compared to most competitors
  • Free-tier limits push serious users toward a paid plan quickly
  • Less distinctive a look than Midjourney at the high end
NightCafe Studio
Image Generation

NightCafe

Under evaluation
  • Friendly hub bundling several underlying models with a big, active community
  • Wide range of styles and an approachable, gamified interface
  • Runs on a credit-based system that can add up for heavy use
  • No single standout model; quality depends on which engine you pick
Sber
Image Generation

Kandinsky 4

Under evaluation
  • Strong regional player with good Russian-language prompt understanding
  • Open-weight variants available for self-hosting
  • Limited adoption and tooling outside its home market
  • Trails the leading Western and Chinese models on general image quality
Luma AI
Image Generation

Luma Photon

Under evaluation
  • Fast, high-quality generation from the same team behind Luma's video models
  • Good for teams already using Luma's video tools, for a consistent pipeline
  • Smaller standalone user base than Midjourney or Flux
  • Less specialized for text-heavy graphics than Ideogram
Freepik
Image Generation

Freepik Mystic

Under evaluation
  • Bundled into Freepik's huge stock-asset library, useful for marketing teams
  • Good style controls tuned for commercial, ready-to-use graphics
  • Less cutting-edge than the pure-play frontier image labs
  • Best value mainly for existing Freepik subscribers
OpenAI
Video Generation

Sora 2

Under evaluation
  • Set the early bar for realistic, physically-consistent video generation
  • Strong brand recognition and broad public familiarity
  • OpenAI closed the Sora app in April 2026 and removes the API on September 24, 2026
  • Superseded by newer models from Google, ByteDance, and Kuaishou
Kuaishou
Video Generation

Kling 3

Under evaluation
  • Excellent character motion and dramatic, controllable camera moves
  • Strong image-to-video animation from a single source frame
  • Frame preservation needs careful checking on longer generations
  • Native audio exists but is often disabled in third-party integrations
Runway
Video Generation

Runway Gen-4.5

Under evaluation
  • Full creative production environment, not just a generation endpoint
  • Strong iteration, editing, and referencing tools for professional workflows
  • Audio still has to be added separately in most pipelines
  • Steeper learning curve than simple prompt-and-generate tools
Luma AI
Video Generation

Luma Ray 3.2

Under evaluation
  • High cinematic motion quality with keyframe-based workflows
  • Good middle ground between simplicity and creative control
  • Less brand visibility than Google, OpenAI, or Kuaishou's offerings
  • Fewer third-party integrations than the bigger platforms
Pika
Video Generation

Pika 2.5

Under evaluation
  • Very fast iteration, with renders in around 40 seconds
  • Distinctive editing features like Pikaswaps and Pikaframes for creative control
  • Prioritizes speed over the maximum quality ceiling of top-tier rivals
  • Sits below the leaderboard-topping models on raw fidelity
Zhipu AI
Video Generation

CogVideoX

Under evaluation
  • Open-weight and self-hostable, popular in the open-source video community
  • Reasonable quality-to-compute ratio for local generation
  • Trails proprietary leaders like Veo or Kling on realism
  • Shorter clip lengths and lower resolution than newer open models
Shengshu AI
Video Generation

Vidu Q3

Under evaluation
  • Native audio-video generation in clips up to 16 seconds
  • Specialized strength in animated-series style production
  • Limited comparative benchmarking outside China-focused coverage
  • Smaller international presence than ByteDance or Kuaishou's models
Genmo AI
Video Generation

Genmo Mochi 1

Under evaluation
  • Was the largest open-weight video model at launch, useful for research
  • Fully open, self-hostable diffusion transformer
  • Capped at 480p and about 5 seconds per clip
  • Superseded by newer open video models
Lightricks
Video Generation

LTX-2.3

Under evaluation
  • Native 4K output at 50fps with synchronized stereo audio, unusually high fidelity for the category
  • Portrait-native training makes it well suited to mobile-first content
  • Larger companies need a commercial license beyond the free tier
  • More resource-intensive to run than lighter open models
Synthesia
Video Generation

Synthesia

Under evaluation
  • Large library of 230+ presenter avatars across 140+ languages
  • Purpose-built for corporate training and internal communications video
  • Not a general-purpose video generator; narrow use case
  • Avatar delivery can feel stiff compared to fully generative video
D-ID
Video Generation

D-ID

Under evaluation
  • Real-time streaming talking avatars with strong API support
  • Well suited to customer-service and interactive-agent use cases
  • Focused on avatars, not general scene or creative video generation
  • Less useful outside conversational or presenter-style formats
Suno
Audio, Voice & Music

Suno v5

Under evaluation
  • Best-in-class vocals, capturing whispers, vibrato, and emotional nuance
  • Full song structure with proper verse, chorus, and bridge arrangement
  • Rap and spoken word still sound noticeably synthetic
  • No official public API yet; a partner API was only announced as exploratory in July 2026
Udio
Audio, Voice & Music

Udio v4

Under evaluation
  • Inpainting lets you regenerate one section without redoing the whole track
  • Stem separation and an official API for paid tiers
  • Public documentation of the v4 release is thin compared with Suno’s
  • Check licensing terms carefully before commercial use of generated music
ElevenLabs
Audio, Voice & Music

ElevenLabs (Voice & Music)

Under evaluation
  • Trained on licensed catalogs, giving strong legal safety for commercial use
  • Realistic voice cloning and text-to-speech with broad API access
  • Music composition quality trails Suno and Udio
  • Generation is slower and pricier than most competitors
Google
Audio, Voice & Music

Google Lyria 3

Under evaluation
  • Generates vocals with auto-written lyrics from text, image, or video prompts
  • High output quality for short-form music
  • Full-length songs need the Lyria 3 Pro tier; the base model makes short clips
  • Google has already moved on to Lyria 3.5 in its own music tools
MiniMax
Audio, Voice & Music

MiniMax Music 2.5

Under evaluation
  • Most affordable API-based music option available
  • Handles niche genre details well, with full commercial rights
  • Third-party API routes often cap clip length well below native limits
  • Much smaller brand recognition than Suno or Udio
Stability AI
Audio, Voice & Music

Stable Audio 2.5

Under evaluation
  • Generates both music and sound effects, useful for production work
  • Audio inpainting for fine-tuning specific sections
  • No vocal generation
  • Commercial use is restricted by revenue thresholds
OpenAI
Audio, Voice & Music

Whisper

Under evaluation
  • Robust, widely-used open speech-to-text across many languages and accents
  • Free and self-hostable, with a large surrounding tool ecosystem
  • No built-in speaker diarization out of the box
  • Architecture is aging relative to newer transcription models
Murf
Audio, Voice & Music

Murf AI

Under evaluation
  • Granular voice control, including emphasis, pitch, pacing, and pronunciation via IPA
  • Bundles voice cloning, dubbing, and translation alongside core text-to-speech
  • Voice library size and language coverage are less clearly documented than rivals
  • Free-tier limits and character caps aren't fully transparent
Play.ht
Audio, Voice & Music

Play.ht (PlayAI)

Under evaluation
  • Large voice library with strong multilingual coverage
  • Good API access for developers building voice into products
  • Voice realism can vary noticeably across less common languages
  • Pricing tiers can get expensive at high usage volumes
WellSaid Labs
Audio, Voice & Music

WellSaid Labs

Under evaluation
  • Studio-quality, natural-sounding voices favored for corporate narration
  • Strong focus on brand-safe, licensed voice talent
  • Smaller voice selection than mass-market competitors
  • Positioned mainly at enterprise budgets, less accessible for casual users
Descript
Audio, Voice & Music

Descript Overdub

Under evaluation
  • Voice cloning built directly into a full audio and video editing workflow
  • Lets you edit spoken audio like text, moving words to reshape a recording
  • Requires recording your own voice samples to train a usable clone
  • Best value comes from Descript's editor, not the voice model alone
AIVA Technologies
Audio, Voice & Music

AIVA

Under evaluation
  • Focused on instrumental composition for film, games, and content creators
  • Lets users guide style and structure with more compositional control than most
  • No vocal generation
  • Less mainstream brand recognition than Suno or Udio
Boomy
Audio, Voice & Music

Boomy

Under evaluation
  • Extremely fast, one-click song creation aimed at total beginners
  • Built-in path to distribute finished tracks to streaming platforms
  • Creative control is shallow compared to Suno or Udio
  • Output quality trails the leading music generators
Perplexity AI
Search & Enterprise

Perplexity Sonar / Deep Research

Under evaluation
  • Real-time web synthesis with inline, checkable citation links
  • Fast, research-focused interface built around multi-query workflows
  • Reasoning depth is shallower than frontier general-purpose models
  • Built for web research rather than analysing large private document sets
Google
Search & Enterprise

NotebookLM

Under evaluation
  • Purpose-built, cited Q&A grounded strictly in your own uploaded sources
  • Generates useful audio-summary overviews from documents
  • Narrow by design; not a general-purpose assistant
  • Answers only from sources in the notebook, so web material has to be imported first
Cohere
Search & Enterprise

Cohere Command A+

Under evaluation
  • Runs efficiently on minimal hardware thanks to a mixture-of-experts design
  • Strong multilingual coverage across 48 languages, tuned for enterprise RAG
  • Built for enterprise retrieval and agents rather than general consumer chat
  • Smaller enterprise track record than the largest cloud AI vendors
Amazon
Search & Enterprise

Amazon Nova

Under evaluation
  • Deeply integrated with AWS Bedrock and the broader AWS stack
  • Competitive cost and speed for multimodal, high-throughput workloads
  • Trails frontier labs on the hardest reasoning benchmarks
  • Most compelling only if you're already committed to AWS
Glean
Search & Enterprise

Glean

Under evaluation
  • Connects to 275+ enterprise apps with near real-time permission enforcement
  • Flexible deployment (SaaS, VPC, or self-hosted) with a choice of 35+ underlying models
  • Needs upfront setup and integration work across many source systems
  • Value depends heavily on how much of the enterprise stack you connect it to
Microsoft
Search & Enterprise

Microsoft 365 Copilot

Under evaluation
  • Deep, seamless integration with Teams, Outlook, and SharePoint
  • Strong at everyday tasks like document drafting, email summarizing, and spreadsheet analysis
  • Works best on data inside Microsoft 365; other systems need connectors or plugins
  • Complex licensing, with multiple SKUs across Copilot, Copilot Studio, and Power Platform
Writer
Search & Enterprise

Writer Palmyra

Under evaluation
  • Built specifically for enterprise content generation with brand and compliance controls
  • Strong governance features aimed at regulated industries like finance and healthcare
  • Narrower general-purpose reasoning than frontier consumer models
  • Smaller mindshare than the major cloud AI vendors