Text models
Chat & completion pricing per 1M tokens, context, and benchmarks.
The catalog includes 2,503 entries with a price. This table shows models with a current price for input or output tokens.
Only models in the top 20 of a platform's usage chart (the ★ badge).Only the smartest N% of scored models, by llminfo IQ percentile.
Compare IQ and price on a chart
Loading the chart.
Compare token prices
Loading prices to compare.
Default order. Select a price for its source and range, or a score for details.
About these columns
- llminfo IQ
- llminfo IQ (capability index). How llminfo IQ is computed
- Coding IQ
- llminfo Coding IQ (coding capability, its own scale). How Coding IQ is computed
- Value
- Value for money. 50 means as expected for its results. Above 50, cheaper per task. How Value for money works
- Yap
- Tokens used per coding task vs. models with the same results. Includes agents that write code. Chattiness on maths doesn't predict it. How the Yap score works
- Epoch Capabilities Index
- Epoch Capabilities Index benchmark details
Showing the first 40 models. The rest are loading.Showing the first 40 models.
| Model | Provider | llminfo IQllminfo IQ (capability index) | Coding IQllminfo Coding IQ (coding capability, its own scale) | Input $/1M | Output $/1M | Max Input Tokens | ValueValue for money. 50 means as expected for its results. Above 50, cheaper per task. | YapTokens used per coding task vs. models with the same results. Includes agents that write code. Chattiness on maths doesn't predict it. | Epoch Capabilities Index |
|---|---|---|---|---|---|---|---|---|---|
Gemini 4 NeonMaker headquartered in US, according to the company's own website. The citation is on the model page.Released 18 Nov 202634 listings | From 88 benchmarks. Plenty of data behind it. | Coding IQ 126.2, 95% interval 95–157. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $3.00 per 1M input tokensAcross 34 comparable listings · $1.50–$5.40 per 1M input tokensCache read · $0.15Batch · $1.50 | Maker's price · Google · $15.00 per 1M output tokensAcross 34 comparable listings · $7.50–$27.00 per 1M output tokensBatch · $7.50 | 200K–1M | Fair value (49.5). Costs per task about what its benchmark results predict. A fair amount of data behind it. | Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | —Not measured | |
Gemini 4 CarbonMaker headquartered in US, according to the company's own website. The citation is on the model page.Released 10 Nov 202637 listings·#6 on NanoGPT this week
#15 on OpenRouter today
#4 in Hacker News this week | From 127 benchmarks. Plenty of data behind it. | Coding IQ 143.6, 95% interval 133–155. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $2.00 per 1M input tokensAcross 37 comparable listings · $1.00–$4.00 per 1M input tokensCache read · $0.10Batch · $1.00 | Maker's price · Google · $10.00 per 1M output tokensAcross 37 comparable listings · $5.00–$20.00 per 1M output tokensBatch · $5.00 | 200K–1M | Good value (72.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (53 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Claude Fable 5.2Maker headquartered in US, according to the company's own website. The citation is on the model page.Released 9 Nov 202660 listings | From 171 benchmarks. Plenty of data behind it. | Coding IQ 146.6, 95% interval 135–158. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $8.00 per 1M input tokensAcross 60 comparable listings · $3.20–$16.00 per 1M input tokensCache read · $0.20Batch · $4.00 | Maker's price · Anthropic · $40.00 per 1M output tokensAcross 60 comparable listings · $16.00–$80.00 per 1M output tokensCache write · $10.00Batch · $20.00 | 200K–1M | Good value (73.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (44 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-6.2 AstraMaker headquartered in US, according to the company's own website. The citation is on the model page.Released 17 Nov 202662 listings | From 120 benchmarks. Plenty of data behind it. | Coding IQ 141.7, 95% interval 133–150. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $8.00 per 1M input tokensAcross 62 comparable listings · $4.00–$24.00 per 1M input tokensCache read · $0.40Batch · $4.00 | Maker's price · OpenAI · $40.00 per 1M output tokensAcross 62 comparable listings · $20.00–$120.00 per 1M output tokensCache write · $10.00Batch · $20.00 | 922K–1.1M | Good value (66.9). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-6 AstraMaker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionGPT-6.2 AstraWe compared release dates and version names in our catalogue.102 listings | From 338 benchmarks. Plenty of data behind it. | Coding IQ 138.9, 95% interval 133–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $10.00 per 1M input tokensAcross 102 comparable listings · $0.6154–$60.00 per 1M input tokensCache read · $1.00Batch · $5.00 | Maker's price · OpenAI · $50.00 per 1M output tokensAcross 102 comparable listings · $3.077–$300.00 per 1M output tokensCache write · $12.50Batch · $25.00 | 272K–1.1M | Good value (62.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (37 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Claude Opus 5.5Maker headquartered in US, according to the company's own website. The citation is on the model page.99 listings·#1 on NanoGPT this week
#5 on OpenRouter this week | From 329 benchmarks. Plenty of data behind it. | Coding IQ 147.1, 95% interval 142–152. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $4.00 per 1M input tokensAcross 99 comparable listings · $0.8999–$8.00 per 1M input tokensCache read · $0.20Batch · $2.00 | Maker's price · Anthropic · $20.00 per 1M output tokensAcross 99 comparable listings · $4.4996–$40.00 per 1M output tokensCache write · $5.00Batch · $10.00 | 200K–1M | Good value (66.8). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (65 ±12). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Gemini 4 HydrogenMaker headquartered in US, according to the company's own website. The citation is on the model page.Released 12 Nov 202638 listings | From 107 benchmarks. Plenty of data behind it. | Coding IQ 125.7, 95% interval 115–137. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $1.50 per 1M input tokensAcross 38 comparable listings · $0.75–$2.70 per 1M input tokensCache read · $0.075Batch · $0.75 | Maker's price · Google · $7.50 per 1M output tokensAcross 38 comparable listings · $3.75–$13.50 per 1M output tokensBatch · $3.75 | 200K–1M | Good value (65.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-6.1 SolMaker headquartered in US, according to the company's own website. The citation is on the model page.92 listings·#7 on OpenRouter this month | From 239 benchmarks. Plenty of data behind it. | Coding IQ 135.0, 95% interval 119–151. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $2.00 per 1M input tokensAcross 92 comparable listings · $0.1231–$12.00 per 1M input tokensCache read · $0.10Batch · $1.00 | Maker's price · OpenAI · $10.00 per 1M output tokensAcross 92 comparable listings · $0.6154–$60.00 per 1M output tokensCache write · $2.50Batch · $5.00 | 272K–1.1M | Good value (68.3). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (51 ±11). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Gemini 4 ArgonMaker headquartered in US, according to the company's own website. The citation is on the model page.55 listings·#7 on NanoGPT this week
#28 on OpenRouter this month | From 217 benchmarks. Plenty of data behind it. | Coding IQ 139.0, 95% interval 130–148. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $2.00 per 1M input tokensAcross 55 comparable listings · $0.60–$4.40 per 1M input tokensCache read · $0.10Batch · $1.00 | Maker's price · Google · $10.00 per 1M output tokensAcross 55 comparable listings · $3.00–$22.00 per 1M output tokensBatch · $5.00 | 200K–1M | Good value (66.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (66 ±9). Generates more tokens than models with similar coding results. Measured at high effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Claude Fable 5.1Maker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.78 listings | From 253 benchmarks. Plenty of data behind it. | Coding IQ 137.9, 95% interval 133–143. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $10.00 per 1M input tokensAcross 78 comparable listings · $3.846–$15.292 per 1M input tokensCache read · $0.25Batch · $5.00 | Maker's price · Anthropic · $50.00 per 1M output tokensAcross 78 comparable listings · $19.23–$76.462 per 1M output tokensCache write · $12.50Batch · $25.00 | 200K–1M | Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (63 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Qwen4 MaxMaker headquartered in CN, according to the company's own website. The citation is on the model page.Released 11 Nov 202640 listings | From 40 benchmarks. Plenty of data behind it. | Coding IQ 124.3, 95% interval 112–137. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Alibaba Cloud · $1.50 per 1M input tokensAcross 40 comparable listings · $0–$1.60 per 1M input tokensCache read · $0.075 | Maker's price · Alibaba Cloud · $4.50 per 1M output tokensAcross 40 comparable listings · $0–$4.80 per 1M output tokensCache write · $1.875 | 984K–1M | Good value (61.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | —Not measured | —Not measured | |
GPT-6.2 SolMaker headquartered in US, according to the company's own website. The citation is on the model page.Listed 5 Nov 202662 listings | From 104 benchmarks. Plenty of data behind it. | Coding IQ 139.0, 95% interval 130–148. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $2.00 per 1M input tokensAcross 62 comparable listings · $1.00–$6.00 per 1M input tokensCache read · $0.10Batch · $1.00 | Maker's price · OpenAI · $10.00 per 1M output tokensAcross 62 comparable listings · $5.00–$30.00 per 1M output tokensCache write · $2.50Batch · $5.00 | 920K–1.1M | Good value (69.3). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | —Not measured | |
Claude Opus 5Maker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.101 listings | From 389 benchmarks. Plenty of data behind it. | Coding IQ 133.8, 95% interval 128–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $5.00 per 1M input tokensAcross 101 comparable listings · $1.17–$10.00 per 1M input tokensCache read · $0.50Batch · $2.50 | Maker's price · Anthropic · $25.00 per 1M output tokensAcross 101 comparable listings · $5.84–$50.00 per 1M output tokensCache write · $6.25Batch · $12.50 | 100K–1M | Good value (64.6). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (61 ±9). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Claude Sonnet 5.5Maker headquartered in US, according to the company's own website. The citation is on the model page.93 listings·#9 on NanoGPT this week
#16 on OpenRouter this month | From 270 benchmarks. Plenty of data behind it. | Coding IQ 140.0, 95% interval 135–145. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $2.00 per 1M input tokensAcross 93 comparable listings · $0.70–$3.2592 per 1M input tokensCache read · $0.10Batch · $1.00 | Maker's price · Anthropic · $10.00 per 1M output tokensAcross 93 comparable listings · $3.50–$16.2962 per 1M output tokensCache write · $2.50Batch · $5.00 | 200K–1M | Good value (66.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (38 ±9). Generates fewer tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Claude Fable 5Maker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionClaude Fable 5.2We compared release dates and version names in our catalogue.86 listings | From 383 benchmarks. Plenty of data behind it. | Coding IQ 132.1, 95% interval 128–137. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $10.00 per 1M input tokensAcross 86 comparable listings · $3.00–$15.292 per 1M input tokensCache read · $1.00Batch · $5.00 | Maker's price · Anthropic · $50.00 per 1M output tokensAcross 86 comparable listings · $17.50–$76.462 per 1M output tokensCache write · $12.50Batch · $25.00 | 200K–1M | Good value (57.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (53 ±10). Generates about as many tokens as models with similar coding results. Measured at default effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Muse WonderMaker headquartered in US, according to the company's own website. The citation is on the model page.12 listings | From 123 benchmarks. Plenty of data behind it. | Coding IQ 134.3, 95% interval 127–142. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Market median price · NanoGPT · $4.00 per 1M input tokensThe middle listing by blended price (3 parts input to 1 output).Across 12 comparable listings · $4.00–$5.25 per 1M input tokensCache read · $0.20 | Market median price · NanoGPT · $20.00 per 1M output tokensThe middle listing by blended price (3 parts input to 1 output).Across 12 comparable listings · $20.00–$26.25 per 1M output tokens | 1M | Good value (54.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (54 ±10). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Kimi K3.1 ProMaker headquartered in CN, according to the company's own website. The citation is on the model page.Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.171 listings·#5 on OpenRouter today
#1 in Hugging Face today | From 161 benchmarks. Plenty of data behind it. | Coding IQ 131.3, 95% interval 121–141. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Moonshot AI · $2.00 per 1M input tokensAcross 171 comparable listings · $0.2403–$5.00 per 1M input tokensCache read · $0.20 | Maker's price · Moonshot AI · $10.00 per 1M output tokensAcross 171 comparable listings · $0.50–$25.00 per 1M output tokensCache write · $2.00 | 262K–1.1M | Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it. | Typical (51 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-6.1 TerraMaker headquartered in US, according to the company's own website. The citation is on the model page.70 listings | From 130 benchmarks. Plenty of data behind it. | Coding IQ 130.4, 95% interval 118–143. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $1.00 per 1M input tokensAcross 70 comparable listings · $0.1557–$2.00 per 1M input tokensCache read · $0.05Batch · $0.50 | Maker's price · OpenAI · $5.00 per 1M output tokensAcross 70 comparable listings · $0.7784–$10.00 per 1M output tokensCache write · $1.25Batch · $2.50 | 272K–1.1M | Good value (54.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (52 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-6 SolMaker headquartered in US, according to the company's own website. The citation is on the model page.84 listings | From 210 benchmarks. Plenty of data behind it. | Coding IQ 129.5, 95% interval 123–136. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $2.00 per 1M input tokensAcross 84 comparable listings · $0.1231–$4.00 per 1M input tokensCache read · $0.20Batch · $1.00 | Maker's price · OpenAI · $10.00 per 1M output tokensAcross 84 comparable listings · $0.6154–$20.00 per 1M output tokensCache write · $2.50Batch · $5.00 | 272K–1.1M | Good value (63.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (48 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Muse Wonder 1.0Maker headquartered in US, according to the company's own website. The citation is on the model page.12 listings | From 49 benchmarks. Plenty of data behind it. | Coding IQ 127.8, 95% interval 116–139. A fair amount of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Meta · $4.00 per 1M input tokensAcross 12 comparable listings · $4.00 per 1M input tokensCache read · $0.20 | Maker's price · Meta · $20.00 per 1M output tokensAcross 12 comparable listings · $20.00 per 1M output tokens | 1M–1M | Good value (59.9). Cheaper per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | —Not measured | |
GPT-5.6 SolMaker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionGPT-6.2 SolWe compared release dates and version names in our catalogue.106 listings | From 504 benchmarks. Plenty of data behind it. | Coding IQ 125.1, 95% interval 120–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $4.00 per 1M input tokensAcross 106 comparable listings · $0.2462–$8.00 per 1M input tokens | Maker's price · OpenAI · $20.00 per 1M output tokensAcross 106 comparable listings · $1.231–$40.00 per 1M output tokens | 262K–1.1M | Good value (59.9). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (41 ±8). Generates fewer tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-5.5 ProMaker headquartered in US, according to the company's own website. The citation is on the model page.28 listings | From 26 benchmarks. Plenty of data behind it. | —Not enough data yet | Maker's price · OpenAI · $30.00 per 1M input tokensAcross 28 comparable listings · $15.00–$31.74 per 1M input tokensBatch · $15.00 | Maker's price · OpenAI · $180.00 per 1M output tokensAcross 28 comparable listings · $90.00–$190.44 per 1M output tokensBatch · $90.00 | 400K–1.1M | Fair value (48.5). Costs per task about what its benchmark results predict. A fair amount of data behind it. | —Not measured | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Muse Spark 1.3Maker headquartered in US, according to the company's own website. The citation is on the model page.23 listings | From 224 benchmarks. Plenty of data behind it. | Coding IQ 125.1, 95% interval 119–132. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Meta · $1.25 per 1M input tokensAcross 23 comparable listings · $1.14–$1.375 per 1M input tokensCache read · $0.15 | Maker's price · Meta · $4.25 per 1M output tokensAcross 23 comparable listings · $4.25–$4.675 per 1M output tokens | 1M–1M | Fair value (52.1). Costs per task about what its benchmark results predict. Plenty of data behind it. | Chatty (59 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
| From 476 benchmarks. Plenty of data behind it. | Coding IQ 121.5, 95% interval 110–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · OpenAI · $5.00 per 1M input tokensAcross 84 comparable listings · $0–$12.50 per 1M input tokensCache read · $0.50Batch · $2.50 | Maker's price · OpenAI · $30.00 per 1M output tokensAcross 84 comparable listings · $0–$75.00 per 1M output tokensBatch · $15.00 | 270K–1.1M | Good value (64.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (28 ±8). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | ||
Claude Opus 4.8Maker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.107 listings | From 478 benchmarks. Plenty of data behind it. | Coding IQ 120.8, 95% interval 116–125. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $5.00 per 1M input tokensAcross 106 comparable listings · $1.17–$10.00 per 1M input tokensCache read · $0.50Batch · $2.50 | Maker's price · Anthropic · $25.00 per 1M output tokensAcross 106 comparable listings · $5.84–$50.00 per 1M output tokensCache write · $6.25Batch · $12.50 | 100K–1M | Good value (52.9). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Terse (40 ±10). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
DeepSeek V4.1 Flash 1017Maker headquartered in CN, according to the company's own website. The citation is on the model page.Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.Listed 21 Nov 2026304 listings·#1 on OpenRouter this week | From 97 benchmarks. Plenty of data behind it. | Coding IQ 126.8, 95% interval 119–135. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · DeepSeek · $0.30 per 1M input tokensAcross 304 comparable listings · $0.0053–$0.75 per 1M input tokensCache read · $0.006 | Maker's price · DeepSeek · $1.20 per 1M output tokensAcross 304 comparable listings · $0.10–$2.40 per 1M output tokens | 256K–1.1M | Good value (71.4). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
LongCat 2.5Maker headquartered in CN, according to the company's own website. The citation is on the model page.Listed 1 Nov 20267 listings | From 19 benchmarks. A fair amount of data behind it. | —Not enough data yet | Paid median price · Kilo · $0.30 per 1M input tokensThe middle of 3 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 7 comparable listings, free ones included · $0–$0.30 per 1M input tokensCache read · $0.006 | Paid median price · Kilo · $1.20 per 1M output tokensThe middle of 3 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 7 comparable listings, free ones included · $0–$1.20 per 1M output tokens | 1M–1M | —Not enough data yet | —Not measured | —Not measured | |
Gemini 3.8 FlashMaker headquartered in US, according to the company's own website. The citation is on the model page.64 listings·#19 on OpenRouter this month | From 325 benchmarks. Plenty of data behind it. | Coding IQ 121.0, 95% interval 113–129. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $0.75 per 1M input tokensAcross 64 comparable listings · $0.3429–$1.65 per 1M input tokensOver 200K tokens · $0.75Cache read · $0.075Batch · $0.375 | Maker's price · Google · $3.75 per 1M output tokensAcross 64 comparable listings · $1.71–$8.25 per 1M output tokensOver 200K tokens · $3.75Batch · $1.875 | 200K–1M | Good value (55.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (64 ±10). Generates more tokens than models with similar coding results. Measured at high effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GPT-5.4 ProMaker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionGPT-5.5 ProWe compared the version names in our catalogue.35 listings | From 28 benchmarks. Plenty of data behind it. | —Not enough data yet | Maker's price · OpenAI · $30.00 per 1M input tokensAcross 35 comparable listings · $2.50–$31.74 per 1M input tokensBatch · $15.00 | Maker's price · OpenAI · $180.00 per 1M output tokensAcross 35 comparable listings · $15.00–$190.44 per 1M output tokensBatch · $90.00 | 4K–1.1M | Good value (64.9). Cheaper per task than its benchmark results predict. A fair amount of data behind it. | —Not measured | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Kimi K3Maker headquartered in CN, according to the company's own website. The citation is on the model page.Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.Newer versionKimi K3.1We compared the version names in our catalogue.218 listings | From 417 benchmarks. Plenty of data behind it. | Coding IQ 123.8, 95% interval 117–131. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Moonshot AI · $3.00 per 1M input tokensAcross 218 comparable listings · $0–$6.00 per 1M input tokensCache read · $0.30 | Maker's price · Moonshot AI · $15.00 per 1M output tokensAcross 218 comparable listings · $0–$30.00 per 1M output tokensCache write · $3.00 | 262K–1.1M | Good value (60.2). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Gemini 3.7 FlashMaker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionGemini 3.8 FlashWe compared release dates and version names in our catalogue.52 listings | From 267 benchmarks. Plenty of data behind it. | Coding IQ 121.3, 95% interval 113–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Google · $0.75 per 1M input tokensAcross 52 comparable listings · $0.3429–$1.65 per 1M input tokensOver 200K tokens · $0.75Cache read · $0.075Batch · $0.375 | Maker's price · Google · $3.75 per 1M output tokensAcross 52 comparable listings · $1.71–$8.25 per 1M output tokensOver 200K tokens · $3.75Batch · $1.875 | 200K–1M | Good value (60.3). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (59 ±12). Generates more tokens than models with similar coding results. Measured at high effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Qwen3.8 MaxMaker headquartered in CN, according to the company's own website. The citation is on the model page.Newer versionQwen4 MaxWe compared release dates and version names in our catalogue.125 listings | From 312 benchmarks. Plenty of data behind it. | Coding IQ 117.2, 95% interval 112–122. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Alibaba Cloud · $2.00 per 1M input tokensAcross 125 comparable listings · $1.00–$3.4286 per 1M input tokensCache read · $0.25 | Maker's price · Alibaba Cloud · $6.00 per 1M output tokensAcross 125 comparable listings · $3.45–$10.286 per 1M output tokensCache write · $2.50 | 246K–1M | Fair value (49.2). Costs per task about what its benchmark results predict. Plenty of data behind it. | Chatty (62 ±9). Generates more tokens than models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Muse Ember 1.0Maker headquartered in US, according to the company's own website. The citation is on the model page.Released 9 Nov 20268 listings | From 3 benchmarks. Little data behind it, so treat it as a rough estimate. | —Not enough data yet | Maker's price · Meta · $0.10 per 1M input tokensAcross 8 comparable listings · $0.10–$0.1143 per 1M input tokensOver 200K tokens · $0.30Cache read · $0.004 | Maker's price · Meta · $0.50 per 1M output tokensAcross 8 comparable listings · $0.50–$0.5714 per 1M output tokensOver 200K tokens · $1.50 | 1M | —Not enough data yet | —Not measured | —Not measured | |
Claude Opus 4.7Maker headquartered in US, according to the company's own website. The citation is on the model page.Newer versionClaude Opus 5.5We compared release dates and version names in our catalogue.103 listings | From 354 benchmarks. Plenty of data behind it. | Coding IQ 118.2, 95% interval 103–134. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · Anthropic · $5.00 per 1M input tokensAcross 103 comparable listings · $1.17–$30.00 per 1M input tokensCache read · $0.50Batch · $2.50 | Maker's price · Anthropic · $25.00 per 1M output tokensAcross 103 comparable listings · $5.84–$150.00 per 1M output tokensCache write · $6.25Batch · $12.50 | 200K–1M | Fair value (49.6). Costs per task about what its benchmark results predict. Plenty of data behind it. | Terse (38 ±12). Generates fewer tokens than models with similar coding results. Measured at default effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
GLM 5.3Maker headquartered in CN, according to the company's own website. The citation is on the model page.Open weights. The weights are publicly available, according to the provider's model list or a public Hugging Face repo.340 listings·#5 on NanoGPT this week
#6 on OpenRouter this month | From 318 benchmarks. Plenty of data behind it. | Coding IQ 121.0, 95% interval 112–130. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Paid median price · Hugging Face · $1.40 per 1M input tokensThe middle of 311 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 313 comparable listings, free ones included · $0–$2.94 per 1M input tokens | Paid median price · Hugging Face · $4.40 per 1M output tokensThe middle of 311 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 313 comparable listings, free ones included · $0–$9.24 per 1M output tokens | 131K–1.3M | Good value (57.7). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (65 ±10). Generates more tokens than models with similar coding results. Measured at max effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Grok 4.6Maker headquartered in US, according to the company's own website. The citation is on the model page.56 listings | From 220 benchmarks. Plenty of data behind it. | Coding IQ 120.1, 95% interval 109–131. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · xAI · $2.00 per 1M input tokensAcross 56 comparable listings · $1.60–$4.00 per 1M input tokensOver 200K tokens · $4.00Cache read · $0.50 | Maker's price · xAI · $6.00 per 1M output tokensAcross 56 comparable listings · $4.50–$12.00 per 1M output tokensOver 200K tokens · $12.00 | 200K–2M | Good value (56.0). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (47 ±10). Generates about as many tokens as models with similar coding results. Measured at high effort. A fair amount of data behind it. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
DeepSeek V4.1 FlashMaker headquartered in CN, according to the company's own website. The citation is on the model page.232 listings·#11 on NanoGPT recently | From 179 benchmarks. Plenty of data behind it. | Coding IQ 124.7, 95% interval 109–140. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Paid median price · Geodd · $0.30 per 1M input tokensThe middle of 231 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 232 comparable listings, free ones included · $0–$0.8202 per 1M input tokens | Paid median price · Geodd · $1.10 per 1M output tokensThe middle of 231 paid listings with both prices. Free listings are left out of the median.The middle listing by blended price (3 parts input to 1 output).Across 232 comparable listings, free ones included · $0–$3.3534 per 1M output tokens | 256K–1.1M | Good value (68.1). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Chatty (60 ±9). Generates more tokens than models with similar coding results. Measured at default effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Grok 4.7Maker headquartered in US, according to the company's own website. The citation is on the model page.56 listings | From 249 benchmarks. Plenty of data behind it. | Coding IQ 117.4, 95% interval 111–124. Plenty of data behind it. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Maker's price · xAI · $2.00 per 1M input tokensAcross 56 comparable listings · $0.1996–$4.00 per 1M input tokensOver 200K tokens · $4.00Cache read · $0.50Batch · $1.60 | Maker's price · xAI · $6.00 per 1M output tokensAcross 56 comparable listings · $0.5988–$12.00 per 1M output tokensOver 200K tokens · $12.00Batch · $4.80 | 500K–524K | Poor value (41.2). Costs more per task than its benchmark results predict. Plenty of data behind it. | Typical (55 ±14). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | Epoch Capabilities IndexSource: Epoch AI Benchmarking Hub API | |
Muse EmberMaker headquartered in US, according to the company's own website. The citation is on the model page.Released 9 Nov 20267 listings | From 32 benchmarks. A fair amount of data behind it. | Coding IQ 119.7, 95% interval 107–132. Little data behind it, so treat it as a rough estimate. Uses the model's best published results. It has its own scale, separate from llminfo IQ. | Market median price · NanoGPT · $0.10 per 1M input tokensThe middle listing by blended price (3 parts input to 1 output).Across 7 comparable listings · $0.10 per 1M input tokensCache read · $0.004 | Market median price · NanoGPT · $0.50 per 1M output tokensThe middle listing by blended price (3 parts input to 1 output).Across 7 comparable listings · $0.50 per 1M output tokens | 1M | Good value (81.8). Cheaper per task than its benchmark results predict. Plenty of data behind it. | Typical (55 ±9). Generates about as many tokens as models with similar coding results. Measured at max effort. Little data behind it, so treat it as a rough estimate. | —Not measured | |
Qwen3.8 Omni FlashMaker headquartered in CN, according to the company's own website. The citation is on the model page.23 listings | From 33 benchmarks. A fair amount of data behind it. | —Not enough data yet | Maker's price · Alibaba Cloud · $0.15 per 1M input tokensAcross 23 comparable listings · $0.1096–$0.30 per 1M input tokensCache read · $0.016 | Maker's price · Alibaba Cloud · $0.47 per 1M output tokensAcross 23 comparable listings · $0.1096–$0.94 per 1M output tokens | 984K–1M | —Not enough data yet | —Not measured | —Not measured |
Only models with a token price are listed here.
Compare IQ and price on a chart
Loading the chart.
Compare token prices
Loading prices to compare.
Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence)