Skip to content

Newest models

Most recent releases, with whatever independent scores exist so far.

Most recent releases, with whatever independent scores exist so far. Source: OpenRouter / Epoch AI. Variants (reasoning effort, batch) are folded into one row.

#ModelCapabilityReasoningCodingContext$/M
301DeepSeek V3.1DeepSeek139.9——164K$0.42
302seed-oss-36b-instruct—71.5———
303nvidia-nemotron-nano-9b-v2—————
304Mistral Medium 3.1Mistral AI · +1 variant———131K$0.80
305GLM 4.5VZ.ai (Zhipu AI)———66K$0.90
306GPT-5OpenAI · +5 variants150.086.2—400K$3.44
307GPT-5 MiniOpenAI · +5 variants145.575.0—400K$0.69
308GPT-5 NanoOpenAI · +5 variants139.469.4—400K$0.14
309qwen3-4b-instruct-2507—45.8———
310Claude Opus 4.1Anthropic · +2 variants144.177.3—200K$30.00
311gpt-oss-120bOpenAI · +2 variants139.975.8—131K$0.070
312gpt-oss-20bOpenAI · +1 variant137.860.8—131K$0.036
313GLM-4.5 ThinkingZ.ai (Zhipu AI)—————
314Codestral 2508Mistral AI · +1 variant———256K$0.45
315Gemini 2.5 Deep ThinkGoogle—————
316Qwen3 Coder 30B A3B InstructAlibaba (Qwen)———262K$0.12
317Qwen3-30B-A3B-Thinking (Jul 2025)Alibaba (Qwen)139.670.1———
318Qwen3-30B-A3B-Instruct (Jul 2025)Alibaba (Qwen)137.455.6———
319Qwen3 30B A3B Instruct 2507Alibaba (Qwen)———262K$0.084
320chutes/GLM-4.5-FP8—————
321Qwen3-235B-A22B-Thinking (Jul 2025)Alibaba (Qwen)143.980.1———
322Qwen3-235B-A22B-Instruct (Jul 2025)Alibaba (Qwen)138.9————
323GLM 4.5Z.ai (Zhipu AI)———131K$1.00
324GLM 4.5 AirZ.ai (Zhipu AI)———131K$0.31
325Qwen3 235B A22B Thinking 2507Alibaba (Qwen)———131K$0.75
326Qwen3 Coder 480B A35BAlibaba (Qwen)———262K$0.47
327Gemini 2.5 Flash LiteGoogle · +1 variant———1.05M$0.17
328UI-TARS 7BByteDance———128K$0.13
329Qwen3 235B A22B Instruct 2507Alibaba (Qwen)———262K$0.15
330Kimi K2 (Jul 2025)Moonshot AI140.1————
331Kimi K2 InstructMoonshot AI—————
332Kimi K2 0711Moonshot AI———131K$1.00
333Grok 4 Heavy (web app)xAI—————
334Grok 4xAI146.587.0———
335UncensoredVenice———128K$0.38
336Hunyuan A13B InstructTencent———131K$0.25
337Morph V3 FastMorph———82K$0.90
338Morph V3 LargeMorph———262K$1.15
339ERNIE 4.5 VL 424B A47BBaidu———123K$0.63
340Grok-3 minixAI140.376.3———
341Mistral Small 3.2Mistral AI131.749.1———
342Mistral Small 3.2 24BMistral AI———256K$0.13
343Gemini 2.5 Flash (Jun 2025)Google140.8————
344Gemini 2.5 Flash-Lite (Jun 2025)Google133.9————
345Gemini 2.5 FlashGoogle · +1 variant———1.05M$0.85
346Gemini 2.5 ProGoogle · +1 variant———1.05M$3.44
347gemini-2.5-flash-lite-preview-06-17 (Default thinking length)Google—————
348MiniMax M1MiniMax———1M$0.85
349MiniMax-M1-80kMiniMax—————
350o3 ProOpenAI147.4——200K$35.00
— means no published score from that source yet · click a model for every benchmark with its source
Best AI for coding →
Models ranked on DeepSWE
Best AI for reasoning →
Ranked on GPQA Diamond
Best AI for math →
Ranked on competition math
Best open-weight models →
Weights you can run yourself
Cheapest capable models →
Lowest blended API price
Best AI coding tools →
Apps & agents by popularity
Best AI image generators →
Apps & models
Best AI for research →
Answer engines & deep research