Models something cheaper beats
127 of the 134 models carrying both an independent score and a standard-delivery price are dominated: at least one other model scores as high or higher and costs less. Each has a page naming which models, under which workload, and what — if anything — you would give up. The undominated remainder is on the value frontier.
Ordered by the largest measured saving. Prices are blended per million tokens on a balanced workload, as of 2026-08-28. A page here exists only while the finding does — when nothing is both better and cheaper, the page stops being built.
| Model | Provider | Intelligence | $/M | Better and cheaper | With a trade | Best saving |
|---|---|---|---|---|---|---|
| Claude Opus 4 | Anthropic | 1364.3 | $30.00 | 37 | 54 | >99% |
| Claude Opus 4.1 | Anthropic | 1417.6 | $30.00 | 25 | 35 | >99% |
| GPT-4 Turbo | OpenAI | 1271.8 | $15.00 | 69 | 48 | >99% |
| GPT-4o (2024-05-13) | OpenAI | 1300.6 | $7.50 | 40 | 67 | >99% |
| o1 | OpenAI | 1365.9 | $26.25 | 24 | 65 | >99% |
| Claude Sonnet 4 | Anthropic | 1338.2 | $6.00 | 18 | 75 | −99% |
| Command R+ (08-2024) | Cohere | 1229 | $4.38 | 107 | 8 | −99% |
| GPT-4o (2024-08-06) | OpenAI | 1282.7 | $4.38 | 33 | 70 | −99% |
| GPT-5.2 | OpenAI | 1416.5 | $4.81 | 8 | 42 | −99% |
| Claude Sonnet 4.5 | Anthropic | 1438 | $6.00 | 11 | 21 | −98% |
| GPT-4.1 | OpenAI | 1382.3 | $3.50 | 9 | 54 | −98% |
| GPT-5 | OpenAI | 1405.6 | $3.44 | 6 | 47 | −98% |
| Mistral Large 2407 | Mistral | 1266.3 | $3.00 | 12 | 85 | −98% |
| Mistral Medium 3.5 | Mistral | 1420.9 | $3.00 | 3 | 38 | −98% |
| o3 | OpenAI | 1409.2 | $3.50 | 7 | 44 | −98% |
| o3 Mini | OpenAI | 1319.2 | $1.93 | 9 | 69 | −98% |
| o3 Mini High | OpenAI | 1336.6 | $1.93 | 7 | 66 | −98% |
| o4 Mini | OpenAI | 1352.5 | $1.93 | 6 | 60 | −97% |
| R1 | DeepSeek | 1372.7 | $1.15 | 47 | 5 | −97% |
| Claude Haiku 4.5 | Anthropic | 1394.9 | $2.00 | 9 | 41 | −96% |
| Claude Opus 4.5 | Anthropic | 1449.8 | $10.00 | 14 | 8 | −96% |
| Claude Opus 4.8 | Anthropic | 1460.9 | $10.00 | 5 | 10 | −96% |
| Gemma 2 27B | 1231.5 | $0.650 | 73 | 0 | −96% | |
| GPT-5.1 | OpenAI | 1441.4 | $3.44 | 3 | 15 | −96% |
| GPT-5.4 Mini | OpenAI | 1412.1 | $1.69 | 1 | 38 | −96% |
| Llama 3.3 70B Instruct | Meta | 1274.9 | $0.710 | 33 | 38 | −96% |
| MiniMax M1 | MiniMax | 1341.9 | $0.963 | 16 | 40 | −96% |
| Qwen2.5 Coder 32B Instruct | Qwen | 1230.2 | $0.745 | 59 | 16 | −96% |
| Qwen3 235B A22B | Qwen | 1366 | $0.796 | 41 | 5 | −96% |
| GLM 4.5V | Z.ai | 1333.6 | $0.900 | 23 | 31 | −95% |
| Nova 2 Lite | Amazon | 1362.2 | $0.850 | 4 | 44 | −95% |
| Qwen3 Max | Qwen | 1412.7 | $1.56 | 17 | 16 | −95% |
| Qwen3 VL 235B A22B Thinking | Qwen | 1400.6 | $1.30 | 18 | 22 | −95% |
| GPT-5 Mini | OpenAI | 1373.4 | $0.688 | 1 | 29 | −94% |
| GPT-5.5 | OpenAI | 1470.9 | $11.25 | 0 | 11 | −94% |
| Grok 4.3 | xAI | 1397.5 | $1.56 | 0 | 43 | −94% |
| Claude Opus 4.7 | Anthropic | 1490.1 | $10.00 | 2 | 1 | −93% |
| Claude Sonnet 4.6 | Anthropic | 1457.6 | $6.00 | 2 | 10 | −93% |
| GLM 5V Turbo | Z.ai | 1436.8 | $1.90 | 2 | 16 | −93% |
| GPT-4.1 Mini | OpenAI | 1340.5 | $0.700 | 5 | 37 | −93% |
| Llama 3.1 70B Instruct | Meta | 1261 | $0.400 | 50 | 6 | −93% |
| Claude 3 Haiku | Anthropic | 1194.8 | $0.500 | 11 | 37 | −92% |
| DeepSeek V3 | DeepSeek | 1332.6 | $0.450 | 32 | 6 | −92% |
| DeepSeek V3 0324 | DeepSeek | 1375 | $0.438 | 5 | 19 | −92% |
| DeepSeek V3.1 | DeepSeek | 1419.1 | $0.825 | 5 | 21 | −92% |
| Gemini 2.5 Flash | 1417.3 | $0.850 | 2 | 26 | −92% | |
| GLM 4.5 | Z.ai | 1429.4 | $1.00 | 11 | 6 | −92% |
| GPT-5.6 Terra | OpenAI | 1446 | $4.50 | 1 | 15 | −92% |
| Inkling | Thinking Machines | 1439.2 | $1.73 | 0 | 13 | −92% |
| MiniMax M2 | MiniMax | 1342.1 | $0.446 | 8 | 26 | −92% |
| Qwen2.5 72B Instruct | Qwen | 1269.1 | $0.370 | 50 | 3 | −92% |
| Qwen3 235B A22B Thinking 2507 | Qwen | 1413.8 | $0.748 | 15 | 12 | −92% |
| R1 0528 | DeepSeek | 1427.9 | $0.912 | 16 | 3 | −92% |
| Qwen3 Next 80B A3B Thinking | Qwen | 1367.5 | $0.412 | 11 | 12 | −91% |
| Qwen3.5-122B-A10B | Qwen | 1417.9 | $0.715 | 1 | 23 | −91% |
| Claude Sonnet 5 | Anthropic | 1442.3 | $4.00 | 2 | 16 | −90% |
| Gemini 2.5 Pro | 1457.3 | $3.44 | 4 | 5 | −90% | |
| GPT-5.4 Nano | OpenAI | 1372.8 | $0.463 | 1 | 18 | −90% |
| GPT-5.6 Sol | OpenAI | 1454.2 | $4.00 | 0 | 12 | −90% |
| Qwen3 VL 235B A22B Instruct | Qwen | 1420.9 | $0.632 | 4 | 8 | −90% |
| DeepSeek V4 Pro 0423 | DeepSeek | 1439.2 | $1.09 | 0 | 9 | −89% |
| GLM 4.6V | Z.ai | 1374.7 | $0.450 | 3 | 15 | −89% |
| MiniMax M2.5 | MiniMax | 1359 | $0.472 | 7 | 15 | −89% |
| Kimi K3 | Moonshot AI | 1476.3 | $6.00 | 1 | 6 | −88% |
| Qwen3.5 397B A17B | Qwen | 1438.3 | $0.877 | 2 | 8 | −88% |
| Qwen3.5-27B | Qwen | 1407.9 | $0.536 | 4 | 14 | −88% |
| Command R (08-2024) | Cohere | 1187.5 | $0.263 | 27 | 5 | −87% |
| Gemini 3.1 Flash Lite Preview | 1414.8 | $0.563 | 0 | 14 | −87% | |
| GPT-5.4 | OpenAI | 1469.6 | $5.63 | 0 | 7 | −87% |
| Mercury 2 | Inception | 1357.8 | $0.375 | 12 | 7 | −87% |
| Trinity Large Thinking | Arcee | 1341.9 | $0.378 | 1 | 20 | −87% |
| Gemini 3.5 Flash Lite | 1436.5 | $0.850 | 1 | 10 | −86% | |
| Grok 4.6 | xAI | 1443.7 | $3.00 | 2 | 14 | −86% |
| Inkling Small | Thinking Machines | 1411.7 | $0.637 | 0 | 22 | −86% |
| Olmo 3 32B Think | Allen AI | 1298.5 | $0.237 | 11 | 11 | −86% |
| Qwen3.5-35B-A3B | Qwen | 1395.6 | $0.500 | 1 | 21 | −86% |
| Qwen3.6 Max Preview | Qwen | 1446.3 | $2.31 | 10 | 1 | −86% |
| Qwen3.8 27B | Qwen | 1440.8 | $0.956 | 0 | 7 | −86% |
| Grok 4.5 | xAI | 1452.3 | $3.00 | 1 | 11 | −85% |
| MiniMax M2.7 | MiniMax | 1405.3 | $0.525 | 4 | 12 | −85% |
| Qwen3.6 Plus | Qwen | 1436.8 | $0.731 | 2 | 8 | −85% |
| Gemini 3.1 Pro Preview | 1479.5 | $4.50 | 2 | 1 | −84% | |
| GPT-5.6 Luna | OpenAI | 1428.5 | $0.450 | 0 | 5 | −84% |
| Qwen3 30B A3B | Qwen | 1316.9 | $0.215 | 14 | 4 | −84% |
| GLM 4.6 | Z.ai | 1439.8 | $0.875 | 4 | 4 | −83% |
| GPT-4o-mini (2024-07-18) | OpenAI | 1286.6 | $0.263 | 1 | 24 | −83% |
| DeepSeek V3.1 Terminus | DeepSeek | 1419.6 | $0.453 | 8 | 3 | −82% |
| Qwen3 Next 80B A3B Instruct | Qwen | 1418.6 | $0.350 | 1 | 9 | −82% |
| DeepSeek V3.2 Exp | DeepSeek | 1424.4 | $0.305 | 7 | 2 | −81% |
| GLM 4.7 | Z.ai | 1435.3 | $0.738 | 2 | 5 | −81% |
| Gemini 3.5 Flash | 1482.6 | $3.38 | 1 | 0 | −79% | |
| Qwen3.8 Max | Qwen | 1481.9 | $3.00 | 0 | 2 | −79% |
| GLM 4.5 Air | Z.ai | 1382.8 | $0.310 | 4 | 7 | −78% |
| GLM 5.1 | Z.ai | 1464.1 | $1.94 | 2 | 3 | −77% |
| Kimi K2.6 | Moonshot AI | 1455 | $1.71 | 0 | 5 | −76% |
| Gemma 3 27B | 1358.3 | $0.172 | 1 | 9 | −74% | |
| DeepSeek V3.2 | DeepSeek | 1424.6 | $0.290 | 1 | 5 | −73% |
| Qwen3 32B | Qwen | 1340.1 | $0.130 | 9 | 2 | −73% |
| MiniMax M3 | MiniMax | 1434.8 | $0.525 | 0 | 3 | −72% |
| GLM 5.3 | Z.ai | 1476.9 | $2.15 | 1 | 1 | −71% |
| GPT-4.1 Nano | OpenAI | 1284.8 | $0.175 | 0 | 18 | −71% |
| Llama 3.2 3B Instruct | Meta | 1109.7 | $0.120 | 9 | 8 | −70% |
| GLM 4.7 Flash | Z.ai | 1352.9 | $0.145 | 8 | 2 | −68% |
| GPT-5 Nano | OpenAI | 1320.3 | $0.138 | 0 | 12 | −68% |
| Muse Spark 1.1 | Meta | 1478.3 | $2.00 | 0 | 1 | −67% |
| GLM 5.2 | Z.ai | 1465.4 | $1.83 | 0 | 2 | −66% |
| Qwen3 235B A22B Instruct 2507 | Qwen | 1419.3 | $0.205 | 5 | 2 | −66% |
| Phi 4 | Microsoft | 1216.8 | $0.087 | 12 | 0 | −64% |
| Step 3.5 Flash | StepFun | 1403.8 | $0.150 | 3 | 2 | −58% |
| Kimi K2.5 | Moonshot AI | 1445.2 | $0.900 | 0 | 4 | −56% |
| Gemma 3 12B | 1334.2 | $0.075 | 0 | 3 | −55% | |
| GLM 5 | Z.ai | 1445.2 | $0.930 | 3 | 1 | −52% |
| Claude Fable 5 | Anthropic | 1494.7 | $20.00 | 2 | 0 | −50% |
| Gemini 3.6 Flash | 1476.5 | $1.50 | 1 | 0 | −50% | |
| Llama 3.2 1B Instruct | Meta | 1054.8 | $0.070 | 6 | 4 | −50% |
| Gemma 3 4B | 1290.8 | $0.063 | 0 | 4 | −49% | |
| gpt-oss-120b | OpenAI | 1365.6 | $0.070 | 2 | 0 | −49% |
| Granite 4.1 8B | IBM | 1291.6 | $0.063 | 3 | 0 | −49% |
| Mistral Small 3 | Mistral | 1233.6 | $0.058 | 3 | 0 | −46% |
| MiMo-V2.5 | Xiaomi | 1427.3 | $0.175 | 0 | 4 | −42% |
| Qwen3.5-Flash | Qwen | 1397.6 | $0.114 | 1 | 2 | −42% |
| Llama 3.1 8B Instruct | Meta | 1186.7 | $0.058 | 3 | 1 | −38% |
| gpt-oss-20b | OpenAI | 1287.8 | $0.055 | 1 | 2 | −35% |
| Gemma 4 26B A4B | 1434.6 | $0.138 | 1 | 1 | −25% | |
| Qwen3.7 Plus | Qwen | 1456.2 | $0.560 | 0 | 1 | −24% |
| Qwen3 30B A3B Instruct 2507 | Qwen | 1384.3 | $0.084 | 1 | 0 | −22% |
| Claude Opus 4.6 | Anthropic | 1502.7 | $10.00 | 1 | 0 | −0% |
Better and cheaper counts replacements that also match or beat the model on context, maximum output, input modes, tool use and extended reasoning. With a trade counts the ones that score higher and cost less but drop one of those; the trade is named on the model's page. Savings are rounded to whole percentages, so a saving shown as >99% is not a free model. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset, while its redistribution terms are unresolved. Models with no independent score are absent from this comparison in both directions — unrated is not a zero.