AI model leaderboard

59 of 116 models on HokAI have a vendor-checked price; 37 have a published SWE-bench Verified score and 53 a GPQA Diamond score.

Ranked by blended price (three input tokens to one output) among GA models that publish at least one score. Benchmark figures are read from each vendor's model card or the evaluator named beside them; speed is cited from Artificial Analysis. Nothing a vendor pays for changes a rank — see how HokAI checks records.

Status: GA Openness: All Price: All Modality: All

85 of 116 models

ModelCreatorBlended $/MSWE-bench VerifiedGPQA DiamondOutput speedContextData termsCheckedFurther analysis
North Micro Vision InstructCohereopen source$0------128KNo training · stated2026-08-14Record · Compare
Phi-4Microsoftopen source$0.084--56.1% vendor · Dec 202426 tok/s cited AA · Jun 202616KNo training · ZDR option2026-08-27Record · Compare
Nemotron 3 Nano 30B A3BNVIDIAopen weights$0.08738.8% vendor · Dec 202573.0% vendor · Dec 2025--1MNo training · stated2026-08-31Record · Compare
Nemotron 3.5 LightningNVIDIAopen source$0.08751.6% vendor · Aug 202675.4% vendor · Aug 2026295 tok/s cited AA · Aug 20261MNo training · stated2026-08-12Record · Compare
Ministral 3 8BMistral AIopen source$0.15--66.8% vendor · Dec 2025--256KNo training · stated2026-08-30Record · Compare
Ministral 3 14BMistral AIopen source$0.2--71.2% vendor · Sep 202692 tok/s cited AA · Sep 2026256KNo training · stated2026-09-07Record · Compare
GLM-5.3-FlashZ.aiopen source$0.237----49 tok/s cited AA · Aug 20261.0Mstated2026-08-27Record · Compare
Hy3Tencentopen source$0.28255% vendor · Apr 202662% vendor · Apr 202695 tok/s cited AA · Jul 2026256KNo training · ZDR option2026-08-19Record · Compare
MAI-Code-1.1-FlashMicrosoftproprietary$0.4572.6% vendor · Aug 2026----256K--2026-08-14Record · Compare
MiniMax M2.7MiniMaxopen weights$0.525----46 tok/s cited AA · Jun 2026205KNo training · stated2026-08-26Record · Compare
MiniMax M3MiniMaxopen weights$0.525------1.0MNo training · stated2026-08-27Record · Compare
Solar Pro 4Upstageproprietary$0.525--89% vendor · Aug 202638 tok/s cited AA524KNo training · stated2026-08-13Record · Compare
DeepSeek-V4-Pro-0813DeepSeekopen source$0.54480.6% cited AA · Aug 202690.1% cited AA · Aug 202680 tok/s cited AA · Aug 20261Mstated2026-08-14Record · Compare
Gemini 3.1 Flash-LiteGoogle DeepMindproprietary$0.563--86.9% vendor · Sep 2026301 tok/s cited AA · Sep 20261.0MNo training · stated2026-09-08Record · Compare
Qwen3.7-PlusAlibaba Cloudproprietary$0.7------1Mstated2026-08-31Record · Compare
Mistral Large 3Mistral AIopen source$0.75--43.9% vendor · Dec 202552 tok/s cited AA · Jun 2026256KNo training · stated2026-08-19Record · Compare
Gemini 3.5 Flash-LiteGoogle DeepMindproprietary$0.85--86.9% vendor · Sep 2026363 tok/s cited AA · Sep 20261.0Mstated2026-09-09Record · Compare
Seed 2.0 ProByteDanceproprietary$0.94576.5% vendor · Feb 202688.9% vendor · Feb 2026--272KNo training · stated2026-08-27Record · Compare
DeepSeek V4DeepSeekopen source$0.9980.6% vendor · Sep 202690.1% vendor · Sep 202658 tok/s cited AA · Sep 20261MTrains on data · stated2026-09-03Record · Compare
Gemini 3.6 FlashGoogle DeepMindproprietary$1.50----198 tok/s cited AA · Sep 20261MNo training · ZDR option2026-09-09Record · Compare
Gemini 3.7 FlashGoogle DeepMindproprietary$1.50----340 tok/s cited AA · Aug 20261.0MNo training · ZDR option2026-08-14Record · Compare
Gemini 3.8 FlashGoogle DeepMindproprietary$1.50----327 tok/s cited AA · Sep 20261.0MNo training · stated2026-09-04Record · Compare
Grok 4.20xAIproprietary$1.5675% vendor · Mar 202688.9% vendor · Mar 2026265 tok/s cited AA · Apr 20262MNo training · ZDR option2026-08-27Record · Compare
Grok 4.3xAIproprietary$1.56----94 tok/s cited AA · Sep 20261M--2026-09-07Record · Compare
Muse Spark 1.3Meta AIproprietary$2.00--94% vendor · Sep 2026182 tok/s cited AA · Sep 20261MNo training · stated2026-09-04Record · Compare
GLM-5.2Z.aiopen source$2.1562.1% vendor · Jun 202680.3% vendor · Jun 2026113 tok/s cited AA · Jun 20261MNo training · stated2026-08-31Record · Compare
GPT-5.6 LunaOpenAIproprietary$2.2560.2% vendor · Jun 202655.3% vendor · Jun 202690 tok/s cited AA · Jul 2026200KNo training · ZDR option2026-08-19Record · Compare
Grok 4.5xAIproprietary$3.00----90 tok/s cited AA · Aug 2026500Kstated2026-09-01Record · Compare
Grok 4.6xAIproprietary$3.00----68 tok/s cited AA · Aug 2026500KNo training · ZDR option2026-08-13Record · Compare
Qwen3.8-2.4T-A95BAlibaba Cloudopen weights$3.00--92.6% vendor · Aug 2026--1.0Mstated2026-08-13Record · Compare
Qwen3.8-MaxAlibaba Cloudproprietary$3.00------1Mstated2026-08-04Record · Compare
Gemini 3.5 FlashGoogleproprietary$3.3878% vendor · May 202690.4% vendor · May 2026289 tok/s cited AA · Jun 20261.0MNo training · stated2026-08-27Record · Compare
Qwen3.7-MaxAlibaba Cloudproprietary$3.75--92.4% vendor · May 2026113 tok/s cited AA · Jun 20261MNo training · stated2026-08-27Record · Compare
GPT-5.4OpenAIproprietary$5.6380% vendor · Mar 202674.8% vendor · Mar 2026--1MNo training · stated2026-08-27Record · Compare
GPT-5.6 TerraOpenAIproprietary$5.6372.3% vendor · Jun 202668.7% vendor · Jun 202685 tok/s cited AA · Jul 2026200KNo training · ZDR option2026-08-19Record · Compare
Claude Sonnet 4.6Anthropicproprietary$6.0079.6% vendor · Feb 202674.1% vendor · Feb 2026--1MNo training · ZDR option2026-07-16Record · Compare
Claude Sonnet 5Anthropicproprietary$6.0082.1% vendor · Jun 202696.2% vendor · Jun 2026--1MNo training · ZDR option2026-08-31Record · Compare
Kimi K3Moonshot AIopen weights$6.0067.5% vendor · Jul 202693.5% vendor · Jul 2026--1.0MNo training · ZDR option2026-07-19Record · Compare
Claude Opus 4.6Anthropicproprietary$10.0080.8% vendor · Feb 202691.3% vendor · Feb 2026--1MNo training · ZDR option2026-07-16Record · Compare
Claude Opus 4.7Anthropicproprietary$10.0087.6% vendor · Apr 202694.2% vendor · Apr 2026--1MNo training · ZDR option2026-07-16Record · Compare
Claude Opus 4.8Anthropicproprietary$10.0088.6% vendor · May 202693.6% vendor · May 202679 tok/s cited AA1MNo training · ZDR option2026-07-16Record · Compare
Claude Opus 5Anthropicproprietary$10.0097% vendor · Jul 202694.1% vendor · Jul 202654 tok/s cited AA · Jul 20261MNo training · ZDR option2026-07-26Record · Compare
FuguSakana AIproprietary$11.25--95.5% vendor · Jun 2026----stated2026-09-01Record · Compare
GPT-5.5OpenAIproprietary$11.2588.7% vendor · Sep 202693.6% vendor · Sep 202661 tok/s cited AA · Sep 20261MNo training · stated2026-07-16Record · Compare
GPT-5.6 SolOpenAIproprietary$11.2578.5% vendor · Jun 202672.1% vendor · Jun 202675 tok/s cited AA · Jul 2026200KNo training · ZDR option2026-08-19Record · Compare
Sakana FuguSakana AIproprietary$11.25--95.5% vendor · Jun 2026--1MNo training · stated2026-08-30Record · Compare
Claude Fable 5Anthropicproprietary$20.0095% vendor · Jun 2026----1MNo training · ZDR option2026-09-01Record · Compare
Claude Fable 5.1Anthropicproprietary$20.00----66 tok/s cited AA · Sep 20261MNo training · stated2026-09-02Record · Compare
GPT-5.5 ProOpenAIproprietary$67.5088.7% vendor · Apr 2026----1MNo training · stated2026-08-19Record · Compare
Cohere Command A+Cohereopen source----76% cited AA · Sep 2026--128KNo training · stated2026-09-08Record · Compare
DiffusionGemma 26B-A4BGoogleopen source----73.2% vendor · Jun 20261008 tok/s cited AA256KNo training · stated2026-08-19Record · Compare
Gemma 4 12BGoogle DeepMindopen source----78.8% vendor · Jun 2026--262KNo training · stated2026-09-01Record · Compare
Gemma 4 31BGoogle DeepMindopen source----85.7% vendor · Apr 2026--262KNo training · stated2026-08-11Record · Compare
Muse GlimmerMeta AIopen source--76% vendor · Aug 202683.5% vendor · Aug 2026--131KNo training · stated2026-08-11Record · Compare
Pulsar 16BMultiverse Computingopen source----71.4% vendor · Jun 20264808 tok/s cited AA1MNo training · stated2026-08-30Record · Compare
Qwen3.6-27BAlibaba Cloud (Qwen Team)open source--77.2% vendor · Apr 202687.8% vendor · Apr 202657 tok/s cited AA262KNo training · stated2026-08-11Record · Compare
Qwen3.8-27BAlibaba Cloud (Qwen Team)open source----89.2% vendor · Aug 2026--262KNo training · stated2026-08-16Record · Compare
Gemini NanoGoogle DeepMindproprietary$0------4KNo training · stated2026-09-01Record · Compare
Ling 3.0 Flash FinInclusionAIopen source$0------262Kstated2026-09-09Record · Compare
Ministral 3 3BMistral AIopen source$0.1------256KNo training · stated2026-08-30Record · Compare
Seed 2.1 TurboByteDanceproprietary$0.825--------stated2026-08-31Record · Compare
Kimi K2.7-CodeMoonshot AIopen weights$1.71------262KTrains on data · stated2026-08-27Record · Compare
Gemini 3 Pro ImageGoogle DeepMindproprietary$4.50------66KNo training · stated2026-09-01Record · Compare
Gemini Omni 1.1 FlashGoogleproprietary$5.50----------2026-09-02Record · Compare
GPT Image 2OpenAIproprietary$13.50--------No training · ZDR option2026-08-30Record · Compare
MAI-Image-2.5Microsoft AIproprietary$15.50------32KNo training · stated2026-09-01Record · Compare
GPT-5.6-CyberOpenAIproprietary$28.13------400KNo training · stated2026-08-12Record · Compare
BosunHanno Labsopen source--------8KNo training · stated2026-08-31Record · Compare
Cohere Parse 5Cohereproprietary--------8KTrains on data · ZDR option2026-09-02Record · Compare
FLUX.2Black Forest Labsopen weights----------No training · stated2026-08-30Record · Compare
Grok Imagine Image 2.0xAIproprietary----------stated2026-08-11Record · Compare
Kling 3.0Kuaishou Technologyproprietary----------No training · stated2026-08-30Record · Compare
LFM2.5-VL-3BLiquid AIopen weights--------33KNo training · stated2026-08-13Record · Compare
LTX-2.5Lightricksopen weights----------stated2026-08-16Record · Compare
MiniMax H3MiniMaxopen weights----------stated2026-08-04Record · Compare
MiniMax Music 3.0MiniMaxopen weights----------stated2026-08-16Record · Compare
Mistral OCR 4Mistral AIproprietary----------No training · stated2026-08-30Record · Compare
Osmosis-Structure-0.6BOsmosisopen weights------------2026-08-11Record · Compare
Recraft V3Recraftproprietary----------No training · stated2026-08-14Record · Compare
Recraft V4.1Recraftproprietary----------No training · stated2026-08-14Record · Compare
Riverflow 2.0 ProSourcefulproprietary--------8KNo training · stated2026-08-30Record · Compare
Seedance 2.5ByteDanceproprietary------------2026-08-09Record · Compare
SeedRealtimeByteDanceproprietary----------stated2026-08-11Record · Compare
Wan-Animate-2Alibaba Cloudopen source----------No training · stated2026-08-26Record · Compare
WeatherNext 3Google DeepMindproprietary----------No training2026-09-07Record · Compare

-- not published by the vendor or its evaluators · * partial figure

Price vs capability

24 of 116 models

Blended price ($ per 1M tokens, 3:1, lower is better) against SWE-bench Verified (% solved, higher is better), GA models with both values
ModelCreatorBlended $/M (3:1)SWE-bench VerifiedFrontier
Nemotron 3.5 LightningNVIDIA$0.087/M51.56%Yes
Nemotron 3 Nano 30B A3BNVIDIA$0.087/M38.76%--
Hy3Tencent$0.282/M55%Yes
MAI-Code-1.1-FlashMicrosoft$0.45/M72.6%Yes
DeepSeek-V4-Pro-0813DeepSeek$0.544/M80.6%Yes
Seed 2.0 ProByteDance$0.945/M76.5%--
DeepSeek V4DeepSeek$0.99/M80.6%--
Grok 4.20xAI$1.56/M75%--
GLM-5.2Z.ai$2.15/M62.1%--
GPT-5.6 LunaOpenAI$2.25/M60.2%--
Gemini 3.5 FlashGoogle$3.38/M78%--
GPT-5.4OpenAI$5.63/M80%--
GPT-5.6 TerraOpenAI$5.63/M72.3%--
Claude Sonnet 5Anthropic$6.00/M82.1%Yes
Claude Sonnet 4.6Anthropic$6.00/M79.6%--
Kimi K3Moonshot AI$6.00/M67.5%--
Claude Opus 5Anthropic$10.00/M97%Yes
Claude Opus 4.8Anthropic$10.00/M88.6%--
Claude Opus 4.7Anthropic$10.00/M87.6%--
Claude Opus 4.6Anthropic$10.00/M80.8%--
GPT-5.5OpenAI$11.25/M88.7%--
GPT-5.6 SolOpenAI$11.25/M78.5%--
Claude Fable 5Anthropic$20.00/M95%--
GPT-5.5 ProOpenAI$67.50/M88.7%--

24 GA models carry both a vendor-checked price and a SWE-bench Verified score. Prices run from $0.087/M (Nemotron 3.5 Lightning) to $67.50/M (GPT-5.5 Pro); the highest score is 97% (Claude Opus 5) at $10.00/M.

Six models are not beaten on both axes by any other: Nemotron 3.5 Lightning, Hy3, MAI-Code-1.1-Flash, DeepSeek-V4-Pro-0813, Claude Sonnet 5 and Claude Opus 5. Every other point is cheaper-or-better on at most one axis than something on that frontier.

Which model fits your constraints? · All models · Compare models · Smart Match