Qwen 3.7 Max
qwen-3.7-maxQwen's flagship reasoning and coding model.
ReasoningToolsVisionCodingLong context
MODELSOne Humiris key. Every production route.
Introducing Humiris UnlimitedUnlimited tokens. One simple plan.Discover Unlimited MODEL PROVIDER
Qwen · Kimi · GLM
Leading reasoning, coding and multimodal models from Alibaba, Moonshot AI and Z.ai.
AVAILABLE THROUGH HUMIRIS
22 production model routes
ReasoningToolsVisionCodingLong contextOpen weights
MODELCONTEXTINPUT / 1MOUTPUT / 1M
Qwen 3.6 Plus
qwen-3.6-plusLong-context multilingual reasoning at efficient cost.
ReasoningToolsCodingLong context
Qwen 3.6 Flash
qwen-3.6-flashFast multimodal Qwen route for interactive products.
ToolsVisionLong context
Qwen 3.5 397B A17B
qwen-3.5-397bOpen mixture-of-experts model for reasoning and code.
ReasoningToolsOpen weightsCoding
Qwen 3 235B A22B
qwen3-235b-a22bLarge open mixture-of-experts model for math, code and agents.
ReasoningToolsOpen weightsCoding
Qwen 3 30B A3B
qwen3-30b-a3bEfficient open MoE route for local reasoning and tool use.
ReasoningToolsOpen weightsCoding
Qwen 2.5 VL 32B
qwen2.5-vl-32b-instructOpen vision-language model for documents, images and visual reasoning.
VisionReasoningOpen weights
Kimi K2.5
kimi-k2.5Multimodal Moonshot model for reasoning, coding and agent workflows.
ReasoningToolsVisionCoding
Kimi K3
kimi-k3Moonshot's flagship model for long-horizon agent work.
ReasoningToolsCodingLong context
Kimi K2.7 Code
kimi-k2.7-codeSpecialized coding model for repository-scale tasks.
ReasoningToolsCodingLong context
Kimi K2.6
kimi-k2.6Balanced agentic reasoning with reliable tool use.
ReasoningToolsCoding
GLM 5.2
glm-5.2Long-horizon reasoning and software engineering model.
ReasoningToolsCodingLong context
GLM 5.1 HighSpeed
glm-5.1-highspeedLow-latency GLM route for interactive agents.
ReasoningToolsCoding
GLM 5 Turbo
glm-5-turboFast general reasoning and multilingual tool use.
ReasoningToolsCoding
GLM 5.1
glm-5.1Agentic engineering model for sustained autonomous software work.
ReasoningToolsCoding
GLM 5
glm-5Open-weight foundation model for long-horizon engineering agents.
ReasoningToolsCodingOpen weights
GLM 5V Turbo
glm-5v-turboMultimodal coding model specialized for visual programming.
ReasoningToolsVisionCoding
GLM 4.7
glm-4.7Stable multi-step reasoning and agentic coding model.
ReasoningToolsCoding
GLM 4.7 Flash
glm-4.7-flashLightweight high-speed GLM route for interactive agent applications.
ReasoningToolsCoding
GLM 4.6V
glm-4.6vVision-language model with native function calling and thinking control.
ReasoningToolsVision
GLM Image
glm-imageOpen image generation model for complex prompts and text rendering.
ImageVisionOpen weights
GLM OCR
glm-ocrDocument parsing and structured information extraction model.
VisionTools
Pricing is shown in USD per one million tokens. Route availability and preview pricing can vary by region and account.
PROVIDER-REPORTED RESULTS
Official benchmark points
Only exact numeric results published by the provider are shown. Models without a comparable public score are left unscored.
Qwen
Qwen 3 235B A22B
GPQA71.1%
LiveCodeBench55.7%
Thinking mode, bf16 checkpoint
Qwen · Qwen3 model card · April 2025Qwen
Qwen 3 30B A3B
GPQA65.8%
LiveCodeBench57.4%
Base Qwen3 thinking-mode comparison, bf16 checkpoint
Qwen · Qwen3 model card · July 2025Kimi
Kimi K2.5
SWE Pro50.7%
SWE Verified76.8%
Terminal 2.050.8%
GPQA87.6%
MMMU-Pro78.5%
BrowseComp60.6%
LiveCodeBench85.0%
Thinking mode; provider evaluation framework
Moonshot AI · Kimi K2.5 model card · February 2026GLM · Z.ai
GLM 5.2
SWE Pro62.1%
Terminal 2.181.0%
Z.ai reported launch configuration
Z.ai · GLM-5.2 · July 2026GLM · Z.ai
GLM 5
SWE Verified77.8%
Terminal 2.056.2%
GPQA86.0%
BrowseComp62.0%
MCP Atlas67.8%
Z.ai reported launch configuration
Z.ai · GLM-5 · February 2026GLM · Z.ai
GLM 4.7
SWE Verified73.8%
Terminal 2.041.0%
GPQA85.7%
BrowseComp52.0%
MCP Atlas52.0%
Z.ai reproduced comparison configuration
Z.ai · GLM-5 comparison · February 2026READY TO ROUTE
Get started