Benchmarks reference

The evidence, filterable.

A dated snapshot of provider claims and independent evaluations. Every claim links to its source. All measurements come from their publishers.

Open the planner

Size buckets follow total parameter count, from under 10B up to the trillion-scale ranges, and each entry also lists its active parameters where published. Closed models with unpublished counts appear as Undisclosed. Scores reflect claims checked through 2026-08-05. Eval setups vary, so compare scores only within the same suite.

Filters

Narrow the evidence set.

67 of 67 models

Provider

Weights

Size class

Input types

Output types

GPT-5.6 Luna
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (max)51
  Artificial Analysis: GPT-5.6 ↗Checked 2026-08-04
- ARC-AGI-2 (max)59.5%
  ARC Prize: GPT-5.6 results ↗Checked 2026-08-04
Fast/affordable tier of the GPT-5.6 family. Parameter count undisclosed.
GPT-5.6 Terra
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (max)55
  Artificial Analysis: GPT-5.6 Terra vs DeepSeek V4 Pro ↗Checked 2026-08-05
- Coding Agent Index (max)77
  Artificial Analysis: GPT-5.6 ↗Checked 2026-08-05
Mid GPT-5.6 hosted tier between Luna and Sol. Parameter count undisclosed.
GPT-5.6 Sol
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Agents' Last Exam53.6
  OpenAI: GPT-5.6 Sol preview ↗Checked 2026-08-04
- Artificial Analysis Intelligence Index (max)59
  Artificial Analysis: GPT-5.6 ↗Checked 2026-08-04
- Coding Agent Index (max, Codex)80
  Artificial Analysis: GPT-5.6 ↗Checked 2026-08-04
Flagship GPT-5.6 model for long-horizon agentic work. Parameter count undisclosed.
GPT-5.4 mini
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
400K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (xhigh)40
  Artificial Analysis: GPT-5.4 mini ↗Checked 2026-08-05
Compact GPT-5.4 generalist for coding and document work. Parameter count undisclosed.
GPT-5.4 nano
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
400K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (xhigh)38
  Artificial Analysis: GPT-5.4 nano vs Gemini 3.1 Pro ↗Checked 2026-08-05
Lowest-cost GPT-5.4 text tier. Parameter count undisclosed.
DeepSeek V4 Flash 0731
DeepSeek
open weights100 to 500B
Params
284B total · 13B active
Context
1M tokens
in · textout · text
- Terminal-Bench 2.182.7
  DeepSeek V4 Flash 0731 model card ↗Checked 2026-08-04
- DeepSWE54.4
  DeepSeek V4 Flash 0731 model card ↗Checked 2026-08-04
- Agents' Last Exam25.2
  DeepSeek V4 Flash 0731 model card ↗Checked 2026-08-04
MIT open weights; MoE with ~13B active. Official post-training release superseding the preview.
DeepSeek V4 Pro
DeepSeek
open weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textout · text
- Artificial Analysis Intelligence Index (max)44
  Artificial Analysis: GPT-5.6 Terra vs DeepSeek V4 Pro ↗Checked 2026-08-05
Open-weight DeepSeek Pro tier for harder coding and research. Parameter count varies by reporting source.
Gemma 4 E4B
Google
open weightsUnder 10B
Params
4B
Context
128K tokens
in · textin · imagein · audioin · videoout · text
- GPQA Diamond58.6%
  Gemma 4 technical report ↗Checked 2026-08-04
- MMMLU76.6%
  Gemma 4 technical report ↗Checked 2026-08-04
Edge/mobile multimodal utility model from the Gemma 4 family.
Gemma 4 26B A4B
Google
open weights10 to 40B
Params
26B total · 3.8B active
Context
256K tokens
in · textin · imagein · videoout · text
- GPQA Diamond82.3%
  Gemma 4 technical report ↗Checked 2026-08-04
- LiveCodeBench v677.1%
  Gemma 4 model overview ↗Checked 2026-08-04
MoE local generalist: ~26B total / ~3.8B active. Fits 24–32 GB consumer GPUs when quantized.
Gemma 4 31B
Google
open weights10 to 40B
Params
31B
Context
256K tokens
in · textin · imagein · videoout · text
- GPQA Diamond84.3%
  Gemma 4 model card ↗Checked 2026-08-04
- MMLU Pro85.2%
  Gemma 4 technical report ↗Checked 2026-08-04
- LiveCodeBench v680.0%
  Gemma 4 technical report ↗Checked 2026-08-04
Dense Gemma 4 flagship for higher-quality local multimodal work.
Gemini 3.1 Pro
Google
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imagein · audioin · videoout · text
- Artificial Analysis Intelligence Index46
  Artificial Analysis: GPT-5.4 mini vs Gemini 3.1 Pro ↗Checked 2026-08-05
Hosted Gemini Pro multimodal tier. Parameter count undisclosed.
Gemini 3.6 Flash
Google
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imagein · audioin · videoout · text
- Artificial Analysis Intelligence Index (high)50
  Artificial Analysis: Gemini 3.6 Flash vs Claude Sonnet 5 ↗Checked 2026-08-05
Fast multimodal Flash generalist on the current Gemini 3.6 line.
Gemini 3.5 Flash-Lite
Google
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imagein · audioin · videoout · text
- Artificial Analysis Intelligence Index36
  Artificial Analysis: Gemini 3.5 Flash-Lite ↗Checked 2026-08-05
Low-cost multimodal Flash-Lite route for high-volume document and vision work.
Claude Sonnet 5
Anthropic
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (max)53
  Artificial Analysis: Anthropic models ↗Checked 2026-08-05
Premium Claude generalist. Parameter count undisclosed.
Claude Haiku 4.5
Anthropic
closed weightsUndisclosed
Params
Undisclosed
Context
200K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index30
  Artificial Analysis: Anthropic models ↗Checked 2026-08-05
Economy Claude route for focused coding and document work.
Claude Fable 5
Anthropic
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index60
  Artificial Analysis: Claude Fable 5 ↗Checked 2026-08-04
- Humanity's Last Exam53%
  Artificial Analysis: Claude Fable 5 ↗Checked 2026-08-04
Anthropic frontier model; parameter count undisclosed. Some evals fall back to Opus 4.8 under safety guardrails.
Claude Opus 4.8
Anthropic
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Terminal-Bench 2.185.0
  DeepSeek V4 Flash 0731 comparison table ↗Checked 2026-08-04
- DeepSWE58.0
  DeepSeek V4 Flash 0731 comparison table ↗Checked 2026-08-04
Prior Anthropic flagship still used as a fallback and comparison baseline.
Claude Opus 5
Anthropic
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (max)61
  Artificial Analysis: Claude Opus 5 ↗Checked 2026-08-05
Current Claude Opus API tier. Parameter count undisclosed.
gpt-oss-20b
OpenAI
open weights10 to 40B
Params
21B total · 3.6B active
Context
128K tokens
in · textout · text
- GPQA Diamond (high, no tools)71.5%
  OpenAI gpt-oss capability findings ↗Checked 2026-08-05
- Artificial Analysis Intelligence Index (high)15
  Artificial Analysis: gpt-oss-120b vs gpt-oss-20b ↗Checked 2026-08-05
Open-weight MoE (~21B total / 3.6B active). Entry-point coding model for 16 GB-class machines.
gpt-oss-120b
OpenAI
open weights100 to 500B
Params
117B total · 5.1B active
Context
128K tokens
in · textout · text
- GPQA Diamond (high, no tools)80.1%
  OpenAI gpt-oss capability findings ↗Checked 2026-08-05
- SWE-Bench Verified (high)62.4%
  OpenAI gpt-oss capability findings ↗Checked 2026-08-05
- Artificial Analysis Intelligence Index (high)24
  Artificial Analysis: gpt-oss-120b vs gpt-oss-20b ↗Checked 2026-08-05
Open-weight MoE (~117B total / 5.1B active). Capacity-class coding route for 80 GB+ machines.
Llama 4 Scout
Meta
open weights100 to 500B
Params
109B total · 17B active
Context
10M tokens
in · textin · imageout · text
- MMLU Pro (instruct)74.3%
  Meta Llama 4 model card ↗Checked 2026-08-04
- GPQA Diamond (instruct)57.2%
  Meta Llama 4 model card ↗Checked 2026-08-04
- MMMU (instruct)69.4%
  Meta Llama 4 announcement ↗Checked 2026-08-04
Open-weight MoE (17B active / 109B total) with a 10M-token context window.
Llama 4 Maverick
Meta
open weights100 to 500B
Params
400B total · 17B active
Context
1M tokens
in · textin · imageout · text
- MMLU Pro (instruct)80.5%
  Meta Llama 4 model card ↗Checked 2026-08-04
- GPQA Diamond (instruct)69.8%
  Meta Llama 4 model card ↗Checked 2026-08-04
Larger Llama 4 MoE (17B active, 128 experts). Total parameter count commonly reported near 400B.
Qwen3.6 27B
Alibaba
open weights10 to 40B
Params
27B
Context
262K tokens
in · textin · imagein · videoout · text
- Artificial Analysis Intelligence Index (reasoning)37
  Artificial Analysis model comparison ↗Checked 2026-08-04
Open-weight mid-size Qwen tier; strong value for local or self-hosted runs.
Qwen 3.7 Max
Alibaba
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textout · text
- Artificial Analysis Intelligence Index46
  Artificial Analysis: Qwen3.7 Max ↗Checked 2026-08-05
Hosted Qwen Max tier. Parameter count undisclosed.
Qwen 3.7 Plus
Alibaba
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imagein · videoout · text
- Artificial Analysis Intelligence Index39
  Artificial Analysis: Qwen3.7 Plus ↗Checked 2026-08-05
Hosted multimodal Qwen Plus tier. Parameter count undisclosed.
Qwen 3.6 Flash
Alibaba
open weights10 to 40B
Params
35B total · 3B active
Context
256K tokens
in · textin · imagein · videoout · text
- GPQA86.0
  Qwen: Qwen3.6-35B-A3B ↗Checked 2026-08-05
Hosted alias for the open Qwen3.6 35B-A3B checkpoint.
Qwen3.6-35B-A3B
Alibaba
open weights10 to 40B
Params
35B total · 3B active
Context
256K tokens
in · textin · imagein · videoout · text
- SWE-bench Verified73.4
  Qwen: Qwen3.6-35B-A3B ↗Checked 2026-08-05
- VideoMME without subtitles82.5
  Qwen: Qwen3.6-35B-A3B ↗Checked 2026-08-05
Open MoE release behind the Qwen3.6 Flash API tier.
Qwen3.8-Max
Alibaba
closed weights2 to 4T
Params
2.4T total · 95B active
Context
Not published
in · textin · imageout · text
- Release announcementAugust 2, 2026
  Qwen: Qwen3.8-Max release ↗Checked 2026-08-05
- Family overviewQwen3.8
  OpenLM: Qwen3.8 ↗Checked 2026-08-05
The first Max-class Qwen with announced open weights, scheduled for release the week after the August 2 launch. 2.4T total with 95B active. This entry flips to open the day the weights land.
Qwen3.7-235B-A22B
Alibaba
open weights100 to 500B
Params
235B total · 22B active
Context
262K tokens
in · textout · text
- LiveCodeBench v679.4%
  Qwen3.7 model card ↗Checked 2026-08-05
Open-weight MoE flagship of the Qwen3.7 line.
Mistral Small 3.3
Mistral
open weights10 to 40B
Params
24B
Context
128K tokens
in · textin · imageout · text
- GPQA Diamond (5-shot CoT)46.13%
  Mistral Small 3.2 model card ↗Checked 2026-08-05
- HumanEval Plus (Pass@5)92.90%
  Mistral Small 3.2 model card ↗Checked 2026-08-05
Dense 24B multimodal open weights. Scores cite the published Small 3.2 instruct card for the same size class.
Gemini 3.5 Flash
Google
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imagein · audioin · videoout · text
- Artificial Analysis Intelligence Index (high)50
  Artificial Analysis: Gemini 3.5 Flash ↗Checked 2026-08-04
- Artificial Analysis Intelligence Index (launch v4.0)55
  Artificial Analysis: Gemini 3.5 Flash launch ↗Checked 2026-08-04
Hosted multimodal Flash tier. Index scores shifted between methodology versions; both dated claims are linked.
Kimi K3
Moonshot AI
open weights2 to 4T
Params
2.8T
Context
1M tokens
in · textout · text
- Artificial Analysis Intelligence Index57
  Santage LLM comparison (citing AA) ↗Checked 2026-08-04
Open-weight long-horizon coding/reasoning model, commonly reported near 2.8T total parameters.
Kimi K2.7 Code
Moonshot AI
open weights1 to 2T
Params
1T total · 32B active
Context
256K tokens
in · textin · imagein · videoout · text
- Kimi Code Bench v262.0
  Moonshot AI: Kimi K2.7 Code model card ↗Checked 2026-08-05
- MCP Atlas76.0
  Moonshot AI: Kimi K2.7 Code model card ↗Checked 2026-08-05
Open coding model with 1T total parameters and 32B active per token.
GLM-5
Z.ai
open weights500B to 1T
Params
744B total · 40B active
Context
200K tokens
in · textout · text
- Artificial Analysis Intelligence Index50
  Artificial Analysis: GLM-5 ↗Checked 2026-08-05
Open-weight MoE (~744B total / 40B active). Useful open frontier baseline beside Kimi.
GLM-5.2
Z.ai
open weights500B to 1T
Params
744B total · 40B active
Context
200K tokens
in · textout · text
- Artificial Analysis Intelligence Index54
  Artificial Analysis: GLM-5.2 ↗Checked 2026-08-05
Successor release on the GLM-5 base with stronger agentic coding results.
Grok 4.3
xAI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (high)38
  Artificial Analysis: Grok 4.3 ↗Checked 2026-08-05
Prior Grok API tier retained for comparison. Parameter count undisclosed.
Grok Composer 2.5
xAI
closed weightsUndisclosed
Params
Undisclosed
Context
Not published
in · textout · text
- Release capabilityLong-running coding
  xAI: Composer 2.5 ↗Checked 2026-08-05
Coding model available through Grok Build. Parameter count undisclosed.
Grok 4.5
xAI
closed weightsUndisclosed
Params
Undisclosed
Context
500K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index (high)54
  Artificial Analysis: Grok 4.5 ↗Checked 2026-08-05
- Artificial Analysis Coding Agent Index (Grok Build)76
  Artificial Analysis: Grok 4.5 launch ↗Checked 2026-08-05
Current xAI flagship reasoning tier. Parameter count undisclosed.
Mistral Large 3
Mistral
open weights500B to 1T
Params
675B total · 41B active
Context
256K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index16
  Artificial Analysis: Mistral Large 3 ↗Checked 2026-08-05
Open-weight MoE (~675B total / 41B active). Current Mistral Large API tier.
Amazon Nova Pro
Amazon
closed weightsUndisclosed
Params
Undisclosed
Context
300K tokens
in · textin · imagein · videoout · text
- MT-Bench (median, Claude 3.7 Sonnet judge)8.5
  AWS: Benchmarking Amazon Nova ↗Checked 2026-08-05
Bedrock multimodal Pro tier with 300K context. Parameter count undisclosed.
Command A
Cohere
open weights100 to 500B
Params
111B
Context
256K tokens
in · textout · text
- Artificial Analysis Intelligence Index8
  Artificial Analysis: Command A ↗Checked 2026-08-05
Cohere enterprise Command tier. Open weights under a non-commercial research license.
MiniMax-M2
MiniMax
open weights100 to 500B
Params
230B total · 10B active
Context
200K tokens
in · textout · text
- Artificial Analysis Intelligence Index28
  Artificial Analysis: MiniMax-M2 ↗Checked 2026-08-05
Open-weight MoE (~230B total / 10B active) aimed at agentic coding.
Doubao Seed Code
ByteDance
closed weightsUndisclosed
Params
Undisclosed
Context
256K tokens
in · textin · imageout · text
- Artificial Analysis Intelligence Index26
  Artificial Analysis: Doubao Seed Code ↗Checked 2026-08-05
ByteDance Seed coding/reasoning API model. Parameter count undisclosed.
ERNIE 4.5 300B A47B
Baidu
open weights100 to 500B
Params
300B total · 47B active
Context
128K tokens
in · textout · text
- Artificial Analysis Intelligence Index9
  Artificial Analysis: ERNIE 4.5 300B A47B ↗Checked 2026-08-05
Open-weight MoE (~300B total / 47B active) from Baidu Ernie.
Inkling
Thinking Machines Lab
open weights500B to 1T
Params
975B total · 41B active
Context
1M tokens
in · textin · imagein · audioout · text
- Humanity's Last Exam (text only)29.7%
  Thinking Machines Lab: Inkling ↗Checked 2026-08-05
- SWE-bench Verified77.6%
  Thinking Machines Lab: Inkling ↗Checked 2026-08-05
Open MoE model with 975B total parameters and 41B active per token.
Inkling-Small Preview
Thinking Machines Lab
closed weights100 to 500B
Params
276B total · 12B active
Context
256K tokens
in · textin · imagein · audioout · text
- GPQA Diamond88.3%
  Thinking Machines Lab: Inkling-Small preview ↗Checked 2026-08-05
Preview checkpoint with 276B total parameters and 12B active. Full weights are pending.
Nemotron 3 Nano Omni
NVIDIA
open weights10 to 40B
Params
30B total · 3B active
Context
256K tokens
in · textin · imagein · audioin · videoout · text
- VideoMME without subtitles (reasoning)72.2
  NVIDIA: Nemotron 3 Nano Omni report ↗Checked 2026-08-05
Open omni model with 30B total parameters and 3B active per token.
Nemotron 3 Super
NVIDIA
open weights100 to 500B
Params
120B total · 12B active
Context
1M tokens
in · textout · text
- HMMT February 202595.4
  NVIDIA: Nemotron 3 Super report ↗Checked 2026-08-05
Open agentic MoE with 120B total parameters and 12B active per token.
Nemotron 3 Ultra
NVIDIA
open weights500B to 1T
Params
550B total · 55B active
Context
1M tokens
in · textout · text
- Terminal-Bench 2.156.4
  NVIDIA: Nemotron 3 Ultra report ↗Checked 2026-08-05
Open Nemotron flagship with 550B total parameters and 55B active per token.
Mistral Small 4
Mistral
open weights100 to 500B
Params
119B total · 6B active
Context
256K tokens
in · textin · imageout · text
- AA-LCR0.72
  Mistral: Mistral Small 4 ↗Checked 2026-08-05
Apache 2.0 MoE with 119B total parameters and 6B active per token.
Phi-4
Microsoft
open weights10 to 40B
Params
14B
Context
16K tokens
in · textout · text
- GPQA56.1
  Microsoft: Phi-4 model card ↗Checked 2026-08-05
Dense 14B open model focused on compact reasoning workloads.
Phi-4 Mini Instruct
Microsoft
open weightsUnder 10B
Params
3.8B
Context
128K tokens
in · textout · text
- MMLU-Pro (zero-shot CoT)52.8
  Microsoft: Phi-4 Mini model card ↗Checked 2026-08-05
Dense 3.8B open model for local text inference.
Phi-4 Multimodal Instruct
Microsoft
open weightsUnder 10B
Params
5.6B
Context
128K tokens
in · textin · imagein · audioout · text
- MathVista testmini62.4
  Microsoft: Phi-4 Multimodal model card ↗Checked 2026-08-05
Open 5.6B model for multimodal understanding.
MiniMax-M3
MiniMax
open weights100 to 500B
Params
428B total · 23B active
Context
1M tokens
in · textin · imagein · videoout · text
- Artificial Analysis Intelligence Index44
  Artificial Analysis: MiniMax-M3 ↗Checked 2026-08-05
- BrowseComp83.5
  MiniMax: M3 ↗Checked 2026-08-05
Open multimodal MoE with 428B total parameters and 23B active per token.
Fugu Ultra
Sakana AI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- SWE-bench Pro73.7%
  Sakana: Fugu release ↗Checked 2026-08-05
- Terminal-Bench 2.182.1%
  Sakana: Fugu release ↗Checked 2026-08-05
- GPQA Diamond95.5%
  Sakana: Fugu release ↗Checked 2026-08-05
Orchestration model that routes hard tasks to a pool of expert models. Scores are Sakana-reported and independent verification is pending.
Fugu
Sakana AI
closed weightsUndisclosed
Params
Undisclosed
Context
1M tokens
in · textin · imageout · text
- SWE-bench Pro59.0%
  Sakana: Fugu release ↗Checked 2026-08-05
- GPQA Diamond95.5%
  Sakana: Fugu release ↗Checked 2026-08-05
Lower-latency orchestration tier that bills at the routed model rate. Scores are Sakana-reported and independent verification is pending.
Seedance 2.0
ByteDance
closed weightsUndisclosed
Params
Undisclosed
Context
Up to 15s video
in · textin · imagein · audioin · videoout · video
- Artificial Analysis T2V with audio1226 Elo
  Artificial Analysis: video leaderboard ↗Checked 2026-08-05
Native multimodal video generation with synchronized audio.
Seedance 2.5
ByteDance
closed weightsUndisclosed
Params
Undisclosed
Context
Enterprise beta
in · textin · imagein · audioin · videoout · video
- Release statusEnterprise beta
  Seedance release tracker ↗Checked 2026-08-05
Public comparative benchmark scores were unavailable at the check date.
MiniMax H3
MiniMax
open weightsUndisclosed
Params
Undisclosed
Context
Up to 15s video
in · textin · imagein · audioin · videoout · video
- Release capability2K · up to 15s
  MiniMax: H3 ↗Checked 2026-08-05
Open Hailuo-line model with native stereo audio generation.
FLUX 3 Image
Black Forest Labs
closed weightsUndisclosed
Params
Undisclosed
Context
Early access
in · textin · imagein · audioin · videoout · image
- Preliminary image evaluationImproved vs FLUX.2
  Black Forest Labs: FLUX 3 ↗Checked 2026-08-05
Early image synthesis and editing surface from the FLUX 3 foundation model.
Veo 3.1
Google
closed weightsUndisclosed
Params
Undisclosed
Context
Up to 8s video
in · textin · imageout · video
- MovieGenBench T2V overall preferenceBest overall
  Google DeepMind: Veo 3.1 ↗Checked 2026-08-05
Hosted video generation with native audio and reference-image control.
Sora 2 Pro
OpenAI
closed weightsUndisclosed
Params
Undisclosed
Context
Up to 25s video
in · textin · imageout · video
- Artificial Analysis Video Arena1183 Elo
  Artificial Analysis: Sora 2 Pro ↗Checked 2026-08-05
Higher-fidelity Sora tier for longer generated clips.
Kling 3.0
Kuaishou
closed weightsUndisclosed
Params
Undisclosed
Context
Up to 15s video
in · textin · imagein · videoout · video
- Artificial Analysis Video Arena (720p Standard)1211 Elo
  Artificial Analysis: Kling 3.0 Standard ↗Checked 2026-08-05
Kling video generation family with native audio and reference controls.
Runway Gen-4.5
Runway
closed weightsUndisclosed
Params
Undisclosed
Context
Up to 10s video
in · textin · imagein · videoout · video
- Artificial Analysis T2V at launch1247 Elo
  Runway: Gen-4.5 ↗Checked 2026-08-05
Runway video model with prompt and reference-based generation controls.
Wan 2.7
Alibaba
closed weightsUndisclosed
Params
Undisclosed
Context
API release
in · textin · imagein · videoout · video
- Artificial Analysis T2V with audio1164 Elo
  Artificial Analysis: video leaderboard ↗Checked 2026-08-05
Current hosted Wan video generation tier with audio output.
Wan 2.2 A14B
Alibaba
open weights10 to 40B
Params
27B total · 14B active
Context
Open checkpoint
in · textin · imageout · video
- Release architecture14B active
  Alibaba: Wan 2.2 ↗Checked 2026-08-05
Open video MoE with separate denoising experts.
Gemini Omni Flash
Google
closed weightsUndisclosed
Params
Undisclosed
Context
API release
in · textin · imageout · video
- Artificial Analysis T2V with audio1244 Elo
  Artificial Analysis: video leaderboard ↗Checked 2026-08-05
Google video generation tier with synchronized audio output.