Skip to content
AI Atlas
Model finder

Find a model

Describe the need; the atlas applies deterministic rules to observed attributes, current prices and current benchmark results and returns the models that satisfy them — sorted by how many criteria they meet, then best benchmark rank, then release date. No composite score, no single winner.

10 models satisfy at least one of the criteriaNo single winner — these are the observed dimensions; the order is criteria met → best rank → release date.

use case long_contextcontext min 1000000license commercial

ModelWhyObserved dimensionsRanks
Z.ai GLM 5.2Open weightsZ.ai (Zhipu AI)
  • licence MIT License allows commercial use
  • context_length 1,048,576 ≥ 1,000,000
753.3B params · 1.05M context · MIT · text · reasoning · tool calling · released 16 Jun 2026 · from $0.60 in / $2 out · 3 providerstau2-bench#1terminal-bench#12livebench-coding#16+12
DeepSeek V4 Flash Vision ExpOpen weightsDeepSeek
  • licence MIT License allows commercial use
  • context_length 1,048,576 ≥ 1,000,000
304.6B params · 1.05M context · MIT · image · text · reasoning · tool calling · released 31 Aug 2026 · from $0.11 in / $0.33 out · 3 providerslivebench-agentic-coding#4livebench-data-analysis#8livebench-2#21+5
DeepSeek V4 Flash (0731)Open weightsDeepSeek
  • licence MIT License allows commercial use
  • context_length 1,310,720 ≥ 1,000,000
304.2B params · 1.31M context · MIT · text · reasoning · tool calling · released 31 Jul 2026 · from $0.04 in / $0.08 out · 4 providerslivebench-data-analysis#9livebench-reasoning#25livebench-language#30+5
Qwen3.8 27BOpen weightsQwen
  • licence Apache License 2.0 allows commercial use
  • context_length 1,000,000 ≥ 1,000,000
27.8B params · 1M context · Apache-2.0 · image · text · video · reasoning · tool calling · released 5 Aug 2026 · from $0.214 in / $2.55 out · 2 providerslivebench-agentic-coding#11livebench-if#16livebench-data-analysis#24+10
GLM 5.3 FlashOpen weightsZ.ai (Zhipu AI)
  • licence MIT License allows commercial use
  • context_length 1,310,720 ≥ 1,000,000
321.3B params · 1.31M context · MIT · image · text · video · reasoning · tool calling · released 25 Aug 2026 · from $0.075 in / $0.25 out · 4 providersartificial-analysis-intelligence-index#12livebench-coding#19livebench-agentic-coding#21+9
Kimi-Linear-48B-A3B-InstructOpen weightsMoonshot AI
  • licence MIT License allows commercial use
  • context_length 1,000,000 ≥ 1,000,000
49.1B params · 1M context · MIT · released 30 Oct 2025 · 0 providersterminal-bench#159ifbench#288tau2-bench#302+3
MiniMax-M1-80kOpen weightsMiniMax
  • licence Apache License 2.0 allows commercial use
  • context_length 1,000,000 ≥ 1,000,000
456.1B params · 1M context · Apache-2.0 · reasoning · released 13 Jun 2025 · 0 providerstau2-bench#178ifbench#188artificial-analysis-intelligence-index#207+3
MiniMax-M1-40kOpen weightsMiniMax
  • licence Apache License 2.0 allows commercial use
  • context_length 1,000,000 ≥ 1,000,000
456.1B params · 1M context · Apache-2.0 · reasoning · released 5 Jun 2025 · 0 providerstau2-bench#186ifbench#196humanitys-last-exam#229+3
DeepSeek V4 Flash 0423Open weightsDeepSeek
  • licence MIT License allows commercial use
  • context_length 1,048,576 ≥ 1,000,000
290.9B params · 1.05M context · MIT · text · reasoning · tool calling · released 22 Apr 2026 · from $0.067 in / $0.134 out · 2 providers—
DeepSeek V4 Pro 0423Open weightsDeepSeek
  • licence MIT License allows commercial use
  • context_length 1,048,576 ≥ 1,000,000
1.6T params · 1.05M context · MIT · text · reasoning · tool calling · released 22 Apr 2026 · from $0.809 in / $1.62 out · 2 providers—

Green chips = criteria satisfied by an observed fact; dashed amber = satisfied by an estimate (hardware fit). Ranks are positions inside each benchmark's primary comparability group.

MethodDeterministic filters over observed attributes, current prices and current benchmark results of canonical models; sorted by number of satisfied criteria, then best benchmark rank, then release date. No composite score. Hardware fit is an estimate. /methodology