Is GPT-4 Turbo a good deal?
Whether anything in this catalogue beats GPT-4 Turbo on both quality and price, and what you give up if it does. A computation on the current catalogue, not an opinion.
12 undominated of 145 · Oct 6, 2026
As of Oct 6, 2026, GPT-4 Turbo is dominated for Balanced on LMArena. Claude Opus 5.5 scores 240.0 higher and costs 47% less, with a covering envelope. 12 of 145 rated, priced standard models are undominated.
Inspect model evidence Compare differences & requirements
Claude Opus 5.5 is both better and cheaper than GPT-4 Turbo: 240.0 points higher on LMArena and 47% less per million tokens, $7.00 cheaper at this mix.
LMArena Elo · higher is better
Scale starts at 1040 Elo
1271.7
Effective $/M · Balanced · lower is better
$15.00/M
LMArena Elo · higher is better
Scale starts at 1040 Elo
1511.7
Effective $/M · Balanced · lower is better
$8.00/M
GPT-4 Turbo takes text, image, returns up to 4,096 tokens from a 128,000-token context, and is listed by 1 seller.
Compared against 145 rated, priced models on this lens: 80 models dominate it and give up nothing, 5 more dominate it but give something up. GPT-4 Turbo scores 1271.7 at $15.00 per million tokens for this mix.
Envelope-safe replacements
Each row scores at least as high, costs no more, and covers this model’s context, output, modalities, tools, and reasoning. A cheaper narrower model is not listed here.
| Model | LMArena | Effective $/M | You save |
|---|---|---|---|
| Claude Opus 5.5 | 1511.7 +240.0 | $8.00/M | 47% |
| Claude Opus 4.6 | 1503.2 +231.5 | $10.00/M | 33% |
| Claude Opus 5 | 1502.1 +230.4 | $10.00/M | 33% |
| Gemini 3.8 Flash | 1497.0 +225.3 | $3.00/M | 80% |
| MiMo-V2.6-Pro | 1491.0 +219.3 | $0.540/M | 96% |
| Muse Spark 1.3 | 1489.9 +218.2 | $2.00/M | 87% |
| Claude Opus 4.7 | 1489.9 +218.2 | $10.00/M | 33% |
| Gemini 3.7 Flash | 1487.0 +215.3 | $3.00/M | 80% |
| Muse Spark 1.2 | 1483.2 +211.5 | $2.00/M | 87% |
| Gemini 3.1 Pro Preview | 1480.2 +208.5 | $4.50/M | 70% |
| Gemini 3.6 Flash | 1479.5 +207.8 | $1.50/M | 90% |
| Muse Spark 1.1 | 1479.2 +207.5 | $2.00/M | 87% |
| Gemini 3.5 Flash | 1477.7 +206.0 | $3.38/M | 78% |
| Kimi K3 | 1475.9 +204.2 | $3.76/M | 75% |
| GLM 5.3 Flash | 1469.6 +197.9 | $0.238/M | 98% |
| GPT-5.5 | 1466.8 +195.1 | $11.25/M | 25% |
| Claude Sonnet 5.5 | 1466.6 +194.9 | $4.00/M | 73% |
| DeepSeek V4.1 Flash | 1462.5 +190.8 | $0.182/M | 99% |
| Claude Opus 4.8 | 1461.0 +189.3 | $10.00/M | 33% |
| Gemini 2.5 Pro | 1457.8 +186.1 | $3.44/M | 77% |
| Claude Sonnet 4.6 | 1457.8 +186.1 | $6.00/M | 60% |
| MiMo-V2.6-Flash | 1456.4 +184.7 | $0.175/M | 99% |
| GPT-5.6 Sol | 1456.3 +184.6 | $8.00/M | 47% |
| Kimi K2.6 | 1455.5 +183.8 | $0.961/M | 94% |
| Qwen3.7 Plus | 1454.6 +182.9 | $0.560/M | 96% |
| GPT-5.4 | 1452.2 +180.5 | $5.63/M | 63% |
| Claude Opus 4.5 | 1451.0 +179.3 | $10.00/M | 33% |
| Grok 4.5 | 1448.1 +176.4 | $3.00/M | 80% |
| GPT-5.6 Terra | 1446.3 +174.6 | $4.50/M | 70% |
| GPT-6.1 Sol | 1445.6 +173.9 | $4.00/M | 73% |
| Kimi K2.5 | 1445.2 +173.5 | $1.20/M | 92% |
| Gemma 4 31B | 1443.4 +171.7 | $0.205/M | 99% |
| Claude Sonnet 5 | 1442.9 +171.2 | $4.00/M | 73% |
| Inkling | 1441.8 +170.1 | $1.73/M | 89% |
| Qwen3.8 27B | 1440.7 +169.0 | $0.562/M | 96% |
| Claude Sonnet 4.5 | 1438.6 +166.9 | $6.00/M | 60% |
| Qwen3.5 397B A17B | 1438.0 +166.3 | $0.878/M | 94% |
| GLM 5V Turbo | 1437.2 +165.5 | $1.90/M | 87% |
| Qwen3.6 Plus | 1436.7 +165.0 | $0.731/M | 95% |
| Gemini 3.5 Flash Lite | 1434.7 +163.0 | $0.850/M | 94% |
| Gemma 4 26B A4B | 1433.9 +162.2 | $0.127/M | 99% |
| MiniMax M3 | 1432.1 +160.4 | $0.485/M | 97% |
| GPT-5.6 Luna | 1431.0 +159.3 | $0.450/M | 97% |
| MiMo-V2.5 | 1427.5 +155.8 | $0.175/M | 99% |
| Grok 4.6 | 1427.4 +155.7 | $3.00/M | 80% |
| GPT-5.1 | 1422.2 +150.5 | $3.44/M | 77% |
| Mistral Medium 3.5 | 1421.2 +149.5 | $3.00/M | 80% |
| Qwen3 VL 235B A22B Instruct | 1420.1 +148.4 | $0.370/M | 98% |
| Qwen3.5-122B-A10B | 1417.1 +145.4 | $0.715/M | 95% |
| Gemini 2.5 Flash | 1417.0 +145.3 | $0.850/M | 94% |
| Gemini 3.1 Flash Lite Preview | 1415.7 +144.0 | $0.563/M | 96% |
| Inkling Small | 1413.7 +142.0 | $0.638/M | 96% |
| GPT-5.2 | 1412.3 +140.6 | $4.81/M | 68% |
| GPT-5.4 Mini | 1411.4 +139.7 | $1.69/M | 89% |
| o3 | 1409.9 +138.2 | $3.50/M | 77% |
| Qwen3.5-27B | 1408.9 +137.2 | $0.536/M | 96% |
| GPT-5 | 1406.1 +134.4 | $3.44/M | 77% |
| Qwen3 VL 235B A22B Thinking | 1400.7 +129.0 | $1.30/M | 91% |
| Grok 4.7 | 1399.9 +128.2 | $3.00/M | 80% |
| Qwen3.5-Flash | 1397.0 +125.3 | $0.114/M | 99% |
| Grok 4.3 | 1396.7 +125.0 | $1.56/M | 90% |
| Claude Haiku 4.5 | 1396.1 +124.4 | $2.00/M | 87% |
| GPT-6 Sol | 1395.4 +123.7 | $4.00/M | 73% |
| Qwen3.5-35B-A3B | 1394.4 +122.7 | $0.355/M | 98% |
| GPT-6 Luna | 1391.5 +119.8 | $0.200/M | 99% |
| GPT-4.1 | 1383.0 +111.3 | $3.50/M | 77% |
| GLM 4.6V | 1376.8 +105.1 | $0.450/M | 97% |
| GPT-5 Mini | 1373.2 +101.5 | $0.688/M | 95% |
| GPT-5.4 Nano | 1372.1 +100.4 | $0.463/M | 97% |
| Nova 2 Lite | 1361.8 +90.1 | $0.850/M | 94% |
| Gemma 3 27B | 1357.8 +86.1 | $0.100/M | 99% |
| o4 Mini | 1353.2 +81.5 | $1.93/M | 87% |
| GPT-4.1 Mini | 1340.4 +68.7 | $0.700/M | 95% |
| Claude Sonnet 4 | 1339.1 +67.4 | $6.00/M | 60% |
| Gemma 3 12B | 1334.2 +62.5 | $0.075/M | 100% |
| GPT-5 Nano | 1320.0 +48.3 | $0.138/M | 99% |
| GPT-4o (2024-05-13) | 1300.4 +28.7 | $7.50/M | 50% |
| GPT-4o-mini (2024-07-18) | 1286.4 +14.7 | $0.263/M | 98% |
| GPT-4.1 Nano | 1284.8 +13.1 | $0.175/M | 99% |
| GPT-4o (2024-08-06) | 1282.5 +10.8 | $4.38/M | 71% |
Higher score, lower price, named losses
Not a drop-in. The loss is why these are not a recommendation.
| Model | LMArena | Effective $/M | You give up |
|---|---|---|---|
| GLM 5.3 | 1471.4 | $1.00/M | image |
| GLM 5.2 | 1470.4 | $1.12/M | image |
| MiMo-V2.5-Pro | 1465.0 | $0.544/M | image |
| GLM 5.1 | 1461.5 | $1.83/M | image |
| Qwen3.6 Max Preview | 1446.8 | $2.31/M | image |
The link keeps the model and mix, using the current catalogue. Monthly spend and switching cost stay in this browser tab and are left out of the link.
Saved decision references
Save the model, workload, catalogue date and capability-preservation rule in this browser. Spend, switching cost and bill contents are not saved. No account or notifications.
Constraint: replacements must preserve the model’s capabilities; any losses remain named trade-offs.
Historical decisions cannot be fully replayed from saved references: past prices, scores and capability evidence are not stored.