AI for Less
50 sources reporting · latest verification 7 min ago

Providers

Compare models across different providers, sorted by Cost Index for a clear view of their relative cost. The index is calculated using the formula below.

Cost Index

2 × +=$0.5= ×1

How do you use AI?

Real AI workloads can vary widely in their input-to-output ratio. Chat is relatively balanced, while RAG, agents, and long-context tasks often process far more input tokens than they generate. Because input and output are priced differently, the same model can rank very differently depending on the workload you choose.

AI Developers

Showing 193 of 987 models

Anthropic AI Developer

Showing 10 of 23 models

  • claude-haiku-5-5×1.4
    $0.1 / $0.5 · Context 1,000,000
  • claude-haiku-4-5×14
    $1 / $5 · Context 200,000×20 / ×13
  • claude-sonnet-5×28
    $2 / $10 · Arena rank #57 · Context 1,000,000×40 / ×25
  • claude-sonnet-5-5×28
    $2 / $10 · Context 1,000,000×40 / ×25
  • claude-sonnet-4×42
    $3 / $15 · Context 200,000×60 / ×38
  • claude-sonnet-4-6×42
    $3 / $15 · Arena rank #44 · Context 1,000,000×60 / ×38
  • claude-opus-5-5×56
    $4 / $20 · Context 1,000,000×80 / ×50
  • claude-opus-4-5×70
    $5 / $25 · Context 200,000×100 / ×63
  • claude-opus-4-6×70
    $5 / $25 · Arena rank #7 · Context 1,000,000×100 / ×63
  • claude-opus-4-7×70
    $5 / $25 · Arena rank #10 · Context 1,000,000×100 / ×63

OpenAI AI Developer

Showing 10 of 173 models

  • gpt-6-luna×1.4
    $0.1 / $0.5 · Context 1,050,000
  • gpt-4o-mini×1.8
    $0.15 / $0.6 · Context 128,000×3.0 / ×1.5
  • gpt-5.6-luna×3.2
    $0.2 / $1.2 · Arena rank #74 · Context 1,050,000
  • gpt-4.1-mini×4.8
    $0.4 / $1.6 · Context 1,047,576×8.0 / ×4.0
  • gpt-5.4-mini×12
    $0.75 / $4.5 · Arena rank #84 · Context 400,000×15 / ×11
  • gpt-4.1×24
    $2 / $8 · Context 1,047,576×40 / ×20
  • gpt-5-search-api×25
    $1.25 / $10×25 / ×25
  • gpt-6.1-sol×28
    $2 / $10 · Context 1,050,000
  • gpt-6-sol×28
    $2 / $10 · Context 1,050,000
  • gpt-4o×30
    $2.5 / $10 · Context 128,000×50 / ×25

Google AI Developer

Showing 10 of 64 models

  • gemini-2.5-flash-lite×1.2
    $0.1 / $0.4 · Context 1,048,576×2.0 / ×1.0
  • gemini-3.1-flash-lite×4.0
    $0.25 / $1.5 · Context 1,048,576×5.0 / ×3.8
  • gemini-2.5-flash×6.2
    $0.3 / $2.5 · Arena rank #149 · Context 1,048,576×6.0 / ×6.3
  • gemini-3.5-flash-lite×6.2
    $0.3 / $2.5 · Arena rank #70 · Context 1,048,576×6.0 / ×6.3
  • gemini-3-flash-preview×8.0
    $0.5 / $3 · Context 1,048,576×10 / ×7.5
  • gemini-3.6-flash×11
    $0.75 / $3.75 · Arena rank #23 · Context 1,048,576×15 / ×9.4
  • gemini-3.7-flash×11
    $0.75 / $3.75 · Arena rank #18 · Context 1,048,576×15 / ×9.4
  • gemini-3.8-flash×11
    $0.75 / $3.75 · Context 1,048,576×15 / ×9.4
  • gemini-robotics-er-2-preview×14
    $1 / $5 · Context 131,072×20 / ×13
  • gemini-3.5-flash×24
    $1.5 / $9 · Arena rank #30 · Context 1,048,576×30 / ×23

Meta AI Developer

Showing 1 of 1 models

  • muse-spark-1.2×14
    $1.25 / $4.25 · Arena rank #11 · Context 1,048,576×25 / ×11

Z.AI AI Developer

Showing 1 of 11 models

Alibaba / Qwen AI Developer

Showing 10 of 63 models

  • qwen3.7-flash×0.4
    $0.03 / $0.13 · Context 1,000,000
  • qwen-turbo×0.6
    $0.05 / $0.2×1.0 / ×0.5
  • qwen-flash×1.0
    $0.05 / $0.4
  • qwen3.5-flash×1.2
    $0.1 / $0.4 · Arena rank #164
  • qwen3.8-flash×1.5
    $0.15 / $0.47 · Context 1,000,000
  • qwen-plus-latest×4.0
    $0.4 / $1.2
  • qwen-plus×4.0
    $0.4 / $1.2 · Context 1,000,000
  • qwen3.6-flash×4.0
    $0.25 / $1.5 · Context 1,000,000
  • qwen3-coder-flash×4.2
    $0.3 / $1.5 · Context 1,000,000
  • qwen3.7-plus×4.8
    $0.4 / $1.6 · Arena rank #66 · Context 1,000,000

Xiaomi MiMo AI Developer

Showing 1 of 5 models

Mistral AI AI Developer

Showing 6 of 21 models

xAI AI Developer

Showing 8 of 11 models

  • grok-build-0.1×8.0
    $1 / $2 · Context 256,000×20 / ×5.0
  • grok-4.20-0309-non-reasoning×10
    $1.25 / $2.5×25 / ×6.3
  • grok-4.20-0309-reasoning×10
    $1.25 / $2.5×25 / ×6.3
  • grok-4.20-multi-agent-0309×10
    $1.25 / $2.5×25 / ×6.3
  • grok-4.3×10
    $1.25 / $2.5 · Arena rank #93 · Context 1,000,000×25 / ×6.3
  • grok-4.5×20
    $2 / $6 · Arena rank #51 · Context 500,000×40 / ×15
  • grok-4.6×20
    $2 / $6 · Arena rank #72 · Context 500,000×40 / ×15
  • grok-4.7×20
    $2 / $6 · Context 500,000×40 / ×15

Cohere AI Developer

Showing 5 of 31 models

OpenRouter Aggregator

Showing 10 of 435 models

DeepInfra Aggregator

Showing 10 of 85 models

Novita AI Aggregator

Showing 10 of 74 models

Hugging Face Inference Providers Aggregator

Showing 3 of 138 models

Fireworks AI Aggregator

Showing 10 of 22 models

Together AI Aggregator

Showing 5 of 19 models

Cloudflare Aggregator

Showing 8 of 89 models

Amazon Bedrock Aggregator

Showing 10 of 58 models

Arena quality & value

Current model quality against the same Cost Index basis used above.

Quality vs cost

Higher Arena rating is better. Arena rank #1 is the top rank.

11001200130014001500×0.5×1×5×10×50×100×250×500Cost Index (log scale)Arena quality rating ↑ betterAmazon Nova 2 Lite Arena #241 · Rating 1335 Cost Index ×6.2 Cheapest at OpenRouterAnthropic: Claude Opus 4.6 Arena #7 · Rating 1498 Cost Index ×70 Cheapest at Anthropic Arena Top 10 #11Anthropic: Claude Opus 4.7 Arena #10 · Rating 1494 Cost Index ×70 Cheapest at Anthropic Arena Top 10 #22Anthropic: Claude Opus 4.8 Arena #40 · Rating 1474 Cost Index ×70 Cheapest at AnthropicAnthropic: Claude Sonnet 4.6 Arena #44 · Rating 1472 Cost Index ×42 Cheapest at AnthropicAnthropic: Claude Sonnet 5 Arena #57 · Rating 1461 Cost Index ×28 Cheapest at Anthropic Arena evaluated claude-sonnet-5-high (High), not the base model. This reviewed representative evaluation is used for Anthropic: Claude Sonnet 5's rank, filters and quality comparisons.c4ai-aya-expanse-32b Arena #310 · Rating 1267 Cost Index ×5.0 Cheapest at CohereClaude Haiku 4 5 20251001 Arena #142 · Rating 1414 Cost Index ×14 Cheapest at AnthropicClaude Opus 4 5 20251101 Arena #47 · Rating 1470 Cost Index ×70 Cheapest at AnthropicClaude Opus 5 Arena #14 · Rating 1489 Cost Index ×70 Cheapest at Anthropic Arena Top 10 #5 Arena has 2 reviewed evaluations for Claude Opus 5. AI for Less conservatively uses the lowest-performing result: claude-opus-5-max (Max). claude-opus-5-high (High): Arena #13, rating 1490. claude-opus-5-max (Max): Arena #14, rating 1489 — used.5Claude Sonnet 4 5 20250929 Arena #69 · Rating 1455 Cost Index ×42 Cheapest at AnthropicCohere: Command R (08-2024) Arena #320 · Rating 1250 Cost Index ×1.8 Cheapest at CohereCohere: Command R+ (08-2024) Arena #304 · Rating 1276 Cost Index ×30 Cheapest at Coherecommand-a-03-2025 Arena #210 · Rating 1354 Cost Index ×30 Cheapest at CohereDeepSeek V4 Flash 0423 (Preview) Arena #103 · Rating 1436 Cost Index ×0.7 Cheapest at DeepInfraDeepSeek-V3 Arena #206 · Rating 1358 Cost Index ×3.1 Cheapest at DeepInfraDeepSeek-V3-0324 Arena #165 · Rating 1395 Cost Index ×2.8 Cheapest at Hugging Face Inference ProvidersDeepSeek-V3.1 Arena #134 · Rating 1417 Cost Index ×2.9 Cheapest at DeepInfraDeepSeek: DeepSeek V3.1 Terminus Arena #139 · Rating 1415 Cost Index ×3.0 Cheapest at Hugging Face Inference ProvidersDeepSeek: DeepSeek V3.2 Arena #122 · Rating 1425 Cost Index ×1.8 Cheapest at DeepInfraDeepSeek: DeepSeek V3.2 Exp Arena #128 · Rating 1421 Cost Index ×1.9 Cheapest at Novita AIDeepSeek: DeepSeek V4 Pro 0423 Arena #62 · Rating 1458 Cost Index ×7.6 Cheapest at OpenRouterDeepSeek: R1 Arena #163 · Rating 1398 Cost Index ×7.8 Cheapest at Hugging Face Inference ProvidersDeepSeek: R1 0528 Arena #129 · Rating 1421 Cost Index ×6.3 Cheapest at DeepInfraGemini 2.5 Flash Arena #149 · Rating 1409 Cost Index ×6.2 Cheapest at DeepInfraGemini 2.5 Pro Arena #86 · Rating 1445 Cost Index ×25 Cheapest at DeepInfraGemini 3.1 Flash-Lite Preview Arena #109 · Rating 1433 Cost Index ×4.0 Cheapest at OpenRouterGemini 3.1 Pro Preview Arena #17 · Rating 1487 Cost Index ×32 Cheapest at Google Arena Top 10 #77Gemini 3.5 Flash Arena #30 · Rating 1476 Cost Index ×24 Cheapest at DeepInfra Arena has 2 reviewed evaluations for Gemini 3.5 Flash. AI for Less conservatively uses the lowest-performing result: gemini-3.5-flash-medium (Medium). gemini-3.5-flash-high (High): Arena #28, rating 1478. gemini-3.5-flash-medium (Medium): Arena #30, rating 1476 — used.Gemini 3.5 Flash-Lite Arena #70 · Rating 1455 Cost Index ×6.2 Cheapest at GoogleGemini 3.6 Flash Arena #23 · Rating 1482 Cost Index ×11 Cheapest at Google Arena evaluated gemini-3.6-flash-high (High), not the base model. This reviewed representative evaluation is used for Gemini 3.6 Flash's rank, filters and quality comparisons.Gemini 3.7 Flash Arena #18 · Rating 1486 Cost Index ×11 Cheapest at Google Arena Top 10 #8 Arena evaluated gemini-3.7-flash-high (High), not the base model. This reviewed representative evaluation is used for Gemini 3.7 Flash's rank, filters and quality comparisons.8Gemma 2 27B Arena #291 · Rating 1289 Cost Index ×3.9 Cheapest at OpenRouterGemma 3 12B Arena #231 · Rating 1342 Cost Index ×0.5 Cheapest at DeepInfraGemma 3 27B Arena #198 · Rating 1365 Cost Index ×0.6 Cheapest at DeepInfraGemma 3 4B Arena #282 · Rating 1303 Cost Index ×0.3 Cheapest at Amazon BedrockGemma 4 31B Arena #75 · Rating 1452 Cost Index ×6.9 Cheapest at CerebrasGLM-4.5 Arena #148 · Rating 1411 Cost Index ×6.8 Cheapest at OpenRouterGLM-4.5-Air Arena #194 · Rating 1373 Cost Index ×2.2 Cheapest at Hugging Face Inference ProvidersGLM-4.6 Arena #124 · Rating 1424 Cost Index ×6.0 Cheapest at Hugging Face Inference ProvidersGLM-4.6V Arena #188 · Rating 1379 Cost Index ×3.0 Cheapest at Novita AIGLM-4.7 Arena #94 · Rating 1441 Cost Index ×5.1 Cheapest at Hugging Face Inference ProvidersGLM-4.7-Flash Arena #199 · Rating 1365 Cost Index ×1.0 Cheapest at CloudflareGLM-5 Arena #63 · Rating 1458 Cost Index ×6.2 Cheapest at OpenRouterGLM-5.1 Arena #56 · Rating 1464 Cost Index ×9.9 Cheapest at OpenRouterGLM-5.2 Arena #33 · Rating 1475 Cost Index ×7.8 Cheapest at Hugging Face Inference Providers Arena evaluated glm-5.2-max (Max), not the base model. This reviewed representative evaluation is used for GLM-5.2's rank, filters and quality comparisons.GLM-5.3 Arena #27 · Rating 1478 Cost Index ×9.9 Cheapest at OpenRouter Arena evaluated glm-5.3-max (Max), not the base model. This reviewed representative evaluation is used for GLM-5.3's rank, filters and quality comparisons.GLM-5V-Turbo Arena #108 · Rating 1433 Cost Index ×13 Cheapest at OpenRoutergpt-3.5-turbo-0125 Arena #333 · Rating 1226 Cost Index ×5.0 Cheapest at OpenAIgpt-3.5-turbo-1106 Arena #343 · Rating 1204 Cost Index ×8.0 Cheapest at OpenAIGPT-4 Turbo (2024-04-09) Arena #251 · Rating 1324 Cost Index ×100 Cheapest at OpenAIGPT-4 Turbo Preview Arena #270 · Rating 1313 Cost Index ×100 Cheapest at OpenAIgpt-4-0314 Arena #294 · Rating 1288 Cost Index ×240 Cheapest at OpenAIgpt-4-0613 Arena #305 · Rating 1276 Cost Index ×240 Cheapest at OpenAIgpt-4.1-2025-04-14 Arena #140 · Rating 1415 Cost Index ×24 Cheapest at OpenAIgpt-4.1-mini-2025-04-14 Arena #185 · Rating 1383 Cost Index ×4.8 Cheapest at OpenAIgpt-4.1-nano-2025-04-14 Arena #255 · Rating 1322 Cost Index ×1.2 Cheapest at OpenAIGPT-4.5 Preview (2025-02-27) Arena #87 · Rating 1445 Cost Index ×600 Cheapest at OpenAIgpt-4o-mini-2024-07-18 Arena #263 · Rating 1318 Cost Index ×1.8 Cheapest at OpenAIgpt-5.2 Arena #104 · Rating 1435 Cost Index ×35 Cheapest at OpenAIgpt-5.3-chat-latest Arena #77 · Rating 1450 Cost Index ×35 Cheapest at OpenAIgpt-5.6-luna Arena #74 · Rating 1452 Cost Index ×1.6 Cheapest at OpenRouter Arena evaluated gpt-5.6-luna-xhigh (xHigh), not the base model. This reviewed representative evaluation is used for gpt-5.6-luna's rank, filters and quality comparisons.gpt-5.6-sol Arena #20 · Rating 1485 Cost Index ×14 Cheapest at OpenRouter Arena Top 10 #9 Arena evaluated gpt-5.6-sol-xhigh (xHigh), not the base model. This reviewed representative evaluation is used for gpt-5.6-sol's rank, filters and quality comparisons.9gpt-5.6-terra Arena #50 · Rating 1466 Cost Index ×16 Cheapest at OpenRouter Arena evaluated gpt-5.6-terra-xhigh (xHigh), not the base model. This reviewed representative evaluation is used for gpt-5.6-terra's rank, filters and quality comparisons.gpt-oss-120b Arena #213 · Rating 1352 Cost Index ×0.5 Cheapest at Hugging Face Inference Providersgpt-oss-20b Arena #265 · Rating 1317 Cost Index ×0.3 Cheapest at OpenRouterGranite-4.2-30B Arena #233 · Rating 1340 Cost Index ×1.9 Cheapest at Hugging Face Inference ProvidersGranite-4.2-3B Arena #290 · Rating 1290 Cost Index ×0.4 Cheapest at Hugging Face Inference ProvidersGranite-4.2-8B Arena #288 · Rating 1290 Cost Index ×0.7 Cheapest at Hugging Face Inference ProvidersGrok 4.5 Arena #51 · Rating 1466 Cost Index ×20 Cheapest at OpenRouterGrok 4.6 Arena #72 · Rating 1454 Cost Index ×20 Cheapest at OpenRouter Arena evaluated grok-4.6-high (High), not the base model. This reviewed representative evaluation is used for Grok 4.6's rank, filters and quality comparisons.Llama 3 8b Instruct Arena #334 · Rating 1224 Cost Index ×2.8 Cheapest at CloudflareMeta: Llama 3.1 70B Instruct Arena #287 · Rating 1293 Cost Index ×2.4 Cheapest at OpenRouterMeta: Llama 3.1 8B Instruct Arena #341 · Rating 1211 Cost Index ×0.2 Cheapest at Hugging Face Inference ProvidersMeta: Llama 3.2 1B Instruct Arena #396 · Rating 1111 Cost Index ×0.5 Cheapest at CloudflareMeta: Llama 3.2 3B Instruct Arena #367 · Rating 1167 Cost Index ×0.9 Cheapest at OpenRouterMeta: Llama 3.3 70B Instruct Arena #262 · Rating 1318 Cost Index ×1.3 Cheapest at Hugging Face Inference ProvidersMeta: Llama 4 Maverick Arena #247 · Rating 1327 Cost Index ×2.1 Cheapest at OpenRouterMeta: Llama 4 Scout Arena #258 · Rating 1322 Cost Index ×0.9 Cheapest at Hugging Face Inference ProvidersMeta: Muse Spark 1.1 Arena #12 · Rating 1491 Cost Index ×14 Cheapest at OpenRouter Arena Top 10 #44Meta: Muse Spark 1.2 Arena #11 · Rating 1492 Cost Index ×14 Cheapest at Meta Arena Top 10 #3 Arena evaluated muse-spark-1.2 (xHigh) (xHigh), not the base model. This reviewed representative evaluation is used for Meta: Muse Spark 1.2's rank, filters and quality comparisons.3MiniMax-M1 Arena #202 · Rating 1364 Cost Index ×6.6 Cheapest at OpenRouterMiniMax-M2 Arena #227 · Rating 1343 Cost Index ×3.6 Cheapest at Amazon BedrockMiniMax-M2.5 Arena #174 · Rating 1391 Cost Index ×3.2 Cheapest at OpenRouterMiniMax-M2.7 Arena #138 · Rating 1415 Cost Index ×2.5 Cheapest at OpenRouterMiniMax-M3 Arena #96 · Rating 1440 Cost Index ×3.3 Cheapest at Hugging Face Inference ProvidersMistral Large 2407 Arena #267 · Rating 1314 Cost Index ×20 Cheapest at OpenRouterMistral Large 4 Arena #115 · Rating 1429 Cost Index ×6.9 Cheapest at Mistral AIMistral Medium 3.5 Arena #117 · Rating 1426 Cost Index ×21 Cheapest at Mistral AIMistral Small 3 Arena #306 · Rating 1274 Cost Index ×0.4 Cheapest at DeepInfraMoonshotAI: Kimi K2.5 Arena #113 · Rating 1430 Cost Index ×6.3 Cheapest at Hugging Face Inference Providers Arena has 2 reviewed evaluations for MoonshotAI: Kimi K2.5. AI for Less conservatively uses the lowest-performing result: kimi-k2.5-instant (Instant). kimi-k2.5-thinking (Thinking): Arena #78, rating 1450. kimi-k2.5-instant (Instant): Arena #113, rating 1430 — used.MoonshotAI: Kimi K2.6 Arena #58 · Rating 1461 Cost Index ×9.4 Cheapest at OpenRouterMoonshotAI: Kimi K3 Arena #16 · Rating 1488 Cost Index ×23 Cheapest at OpenRouter Arena Top 10 #6 Arena evaluated kimi-k3-max (Max), not the base model. This reviewed representative evaluation is used for MoonshotAI: Kimi K3's rank, filters and quality comparisons.6Nemotron-3-Nano-30B-A3B Arena #269 · Rating 1314 Cost Index ×0.6 Cheapest at Novita AI Arena evaluated nvidia-nemotron-3-nano-30b-a3b-bf16 (BF16), not the base model. This reviewed representative evaluation is used for Nemotron-3-Nano-30B-A3B's rank, filters and quality comparisons.Nvidia Nemotron 3 Ultra 550b A55b Nvfp4 Arena #120 · Rating 1426 Cost Index ×7.2 Cheapest at Hugging Face Inference ProvidersNVIDIA-Nemotron-3-Super-120B-A12B Arena #204 · Rating 1360 Cost Index ×1.1 Cheapest at DeepInfrao1-2024-12-17 Arena #156 · Rating 1402 Cost Index ×180 Cheapest at OpenAIo1-mini Arena #235 · Rating 1337 Cost Index ×13 Cheapest at OpenAIo1-preview Arena #176 · Rating 1389 Cost Index ×180 Cheapest at OpenAIo3-mini Arena #218 · Rating 1348 Cost Index ×13 Cheapest at OpenAIo4-mini-2025-04-16 Arena #172 · Rating 1391 Cost Index ×13 Cheapest at OpenAIOpenAI: GPT-4o (2024-05-13) Arena #223 · Rating 1346 Cost Index ×50 Cheapest at OpenAIOpenAI: GPT-4o (2024-08-06) Arena #239 · Rating 1336 Cost Index ×30 Cheapest at OpenAIOpenAI: GPT-5 Arena #105 · Rating 1435 Cost Index ×25 Cheapest at OpenAI Arena evaluated gpt-5-high (High), not the base model. This reviewed representative evaluation is used for OpenAI: GPT-5's rank, filters and quality comparisons.OpenAI: GPT-5 Mini Arena #175 · Rating 1390 Cost Index ×5.0 Cheapest at OpenAI Arena evaluated gpt-5-mini-high (High), not the base model. This reviewed representative evaluation is used for OpenAI: GPT-5 Mini's rank, filters and quality comparisons.OpenAI: GPT-5 Nano Arena #234 · Rating 1337 Cost Index ×1.0 Cheapest at OpenAI Arena evaluated gpt-5-nano-high (High), not the base model. This reviewed representative evaluation is used for OpenAI: GPT-5 Nano's rank, filters and quality comparisons.OpenAI: GPT-5.1 Arena #98 · Rating 1439 Cost Index ×25 Cheapest at OpenAIOpenAI: GPT-5.4 Arena #53 · Rating 1465 Cost Index ×40 Cheapest at OpenAIOpenAI: GPT-5.4 Mini Arena #84 · Rating 1447 Cost Index ×12 Cheapest at OpenAI Arena evaluated gpt-5.4-mini-high (High), not the base model. This reviewed representative evaluation is used for OpenAI: GPT-5.4 Mini's rank, filters and quality comparisons.OpenAI: GPT-5.4 Nano Arena #159 · Rating 1401 Cost Index ×3.3 Cheapest at OpenAI Arena evaluated gpt-5.4-nano-high (High), not the base model. This reviewed representative evaluation is used for OpenAI: GPT-5.4 Nano's rank, filters and quality comparisons.OpenAI: GPT-5.5 Arena #29 · Rating 1477 Cost Index ×80 Cheapest at OpenAIPhi-4 Arena #318 · Rating 1256 Cost Index ×0.6 Cheapest at DeepInfraQwen: Qwen3 235B A22B Arena #190 · Rating 1375 Cost Index ×2.0 Cheapest at Hugging Face Inference ProvidersQwen: Qwen3 235B A22B Thinking 2507 Arena #161 · Rating 1400 Cost Index ×5.5 Cheapest at Alibaba / QwenQwen: Qwen3 30B A3B Arena #248 · Rating 1327 Cost Index ×1.5 Cheapest at DeepInfraQwen: Qwen3 30B A3B Instruct 2507 Arena #186 · Rating 1382 Cost Index ×1.0 Cheapest at OpenRouterQwen: Qwen3 32B Arena #221 · Rating 1347 Cost Index ×0.8 Cheapest at Hugging Face Inference ProvidersQwen: Qwen3 Next 80B A3B Instruct Arena #162 · Rating 1399 Cost Index ×2.6 Cheapest at DeepInfraQwen: Qwen3 Next 80B A3B Thinking Arena #196 · Rating 1369 Cost Index ×3.0 Cheapest at Alibaba / QwenQwen: Qwen3 VL 235B A22B Instruct Arena #144 · Rating 1413 Cost Index ×2.6 Cheapest at DeepInfraQwen: Qwen3 VL 235B A22B Thinking Arena #166 · Rating 1395 Cost Index ×9.6 Cheapest at OpenRouterQwen: Qwen3.5 397B A17B Arena #92 · Rating 1442 Cost Index ×7.8 Cheapest at DeepInfraQwen: Qwen3.5-122B-A10B Arena #135 · Rating 1416 Cost Index ×5.2 Cheapest at OpenRouterQwen: Qwen3.5-27B Arena #151 · Rating 1409 Cost Index ×6.0 Cheapest at Alibaba / QwenQwen: Qwen3.5-35B-A3B Arena #167 · Rating 1394 Cost Index ×1.8 Cheapest at OpenRouterQwen: Qwen3.6 Max Preview Arena #59 · Rating 1460 Cost Index ×16 Cheapest at OpenRouterQwen: Qwen3.6 Plus Arena #88 · Rating 1443 Cost Index ×5.2 Cheapest at OpenRouterQwen: Qwen3.7 Plus Arena #66 · Rating 1456 Cost Index ×3.8 Cheapest at OpenRouterQwen: Qwen3.8 27B Arena #99 · Rating 1438 Cost Index ×0.0 Cheapest at OpenRouterQwen: Qwen3.8 Max Arena #22 · Rating 1483 Cost Index ×17 Cheapest at DeepInfra Arena Top 10 #1010Qwen2.5 Coder 32b Instruct Arena #309 · Rating 1271 Cost Index ×0.6 Cheapest at Hugging Face Inference ProvidersQwen2.5-72B-Instruct Arena #283 · Rating 1303 Cost Index ×2.2 Cheapest at DeepInfraQwen3-235B-A22B-Instruct-2507 Arena #127 · Rating 1422 Cost Index ×1.5 Cheapest at DeepInfraQwen3-Coder-480B-A35B-Instruct Arena #178 · Rating 1388 Cost Index ×4.6 Cheapest at Hugging Face Inference Providersqwen3-max-2025-09-23 Arena #125 · Rating 1423 Cost Index ×17 Cheapest at Alibaba / Qwenqwen3-max-preview Arena #106 · Rating 1434 Cost Index ×17 Cheapest at Alibaba / Qwenqwen3.5-flash Arena #164 · Rating 1396 Cost Index ×1.2 Cheapest at Alibaba / Qwenqwen3.7-max-preview Arena #36 · Rating 1475 Cost Index ×25 Cheapest at Alibaba / QwenQwq 32b Arena #238 · Rating 1336 Cost Index ×4.6 Cheapest at CloudflareTencent: Hy3 Arena #73 · Rating 1454 Cost Index ×1.6 Cheapest at Hugging Face Inference ProvidersxAI: Grok 4.3 Arena #93 · Rating 1442 Cost Index ×10 Cheapest at OpenRouterXiaomi: MiMo-V2.5 Arena #107 · Rating 1434 Cost Index ×1.1 Cheapest at Hugging Face Inference ProvidersXiaomi: MiMo-V2.5-Pro Arena #48 · Rating 1468 Cost Index ×3.5 Cheapest at OpenRouterXiaomi: MiMo-V2.6-Flash Arena #76 · Rating 1451 Cost Index ×1.1 Cheapest at Novita AIXiaomi: MiMo-V2.6-Pro Arena #26 · Rating 1480 Cost Index ×3.5 Cheapest at Novita AIZ.ai: GLM 4.5V Arena #212 · Rating 1352 Cost Index ×6.0 Cheapest at Hugging Face Inference ProvidersZ.ai: GLM 5.3 Flash Arena #38 · Rating 1475 Cost Index ×1.6 Cheapest at Cloudflare

All Arena-ranked models Arena Top 10

Arena Top 10

The highest-ranked Arena models tracked by AI for Less, shown with their current Cost Index.

  1. Anthropic: Claude Opus 4.6 Arena #7 · Cost ×70
  2. Anthropic: Claude Opus 4.7 Arena #10 · Cost ×70
  3. Meta: Muse Spark 1.2 Arena #11 · Cost ×14
  4. Meta: Muse Spark 1.1 Arena #12 · Cost ×14
  5. Claude Opus 5 Arena #14 · Cost ×70
  6. MoonshotAI: Kimi K3 Arena #16 · Cost ×23
  7. Gemini 3.1 Pro Preview Arena #17 · Cost ×32
  8. Gemini 3.7 Flash Arena #18 · Cost ×11
  9. gpt-5.6-sol Arena #20 · Cost ×14
  10. Qwen: Qwen3.8 Max Arena #22 · Cost ×17

Quality data: Arena leaderboard dataset · CC BY 4.0 · Text leaderboard · style-controlled · published 2026-10-08