Time per task: how long one benchmark task takes
How long each configuration took per task of Artificial Analysis' Intelligence Index, averaged over the whole run. Most tasks are agentic and run many turns, so this is minutes of work, not the wait for one reply; more reasoning effort usually means a longer task.
As of 13 Nov 2026, BenchLeader ranks GPT-6 Luna first on the Time per task: how long one benchmark task takes board at 37 s, ahead of Gemini 3.1 Flash Lite (53 s) and GPT-6.1 Terra (70 s).
120 models · data as of 13 Nov 2026
Quality against what it costs to get there
Hover or tap a model to pick it out and see its other effort settings; click to pin it (Esc clears). The eye hides a model.
Time per task is Artificial Analysis’ own figure for each configuration’s Intelligence Index run, which it estimates from the output tokens and the model’s output speed. Most tasks are agentic and run many turns, so a task takes minutes, not the seconds one reply does. Each dot is a model at its best reasoning-effort setting (highest Index); pick a model out to see its other settings, joined lowest effort first; up and to the left is better.
How far the effort dial moves each model
One bar per model, from its lowest to its highest figure across its reasoning-effort settings, a mark at each setting. The same models as the chart above.
Tokens per task
output tokens, reasoning includedOutput tokens, full Index run
every task of the run, as AA totals itArtificial Analysis’ own total for the whole run: nine evaluations and over 5,000 tasks, several run three or five times. Its per-task figures are averages weighted by each evaluation’s share of the index, so they do not multiply back to this.
Cost per task
USD at list priceCost of a full Index run
USD at list price, as AA totals itArtificial Analysis’ own total for the whole run: nine evaluations and over 5,000 tasks, several run three or five times. Its per-task figures are averages weighted by each evaluation’s share of the index, so they do not multiply back to this.
Time per task
per Index taskCost, tokens and time per task, and the full-run totals, are from Artificial Analysis’ Intelligence Index runs, shown with credit and not offered for reuse. The Index is BenchLeader’s.
- 1GPT-6 Luna (low)37 s
- 2Gemini 3.1 Flash Lite53 s
- 3GPT-6.1 Terra (low)70 s
- 4Gemini 4 Hydrogen (low)80 s
- 5Mistral Small 4 (thinking)1.7 min
- 6Gemini 3.5 Flash Lite1.7 min
- 7GPT-6.1 Sol (low)2 min
- 8Beam 1.02.1 min
- 9Claude Haiku 5.5 (low)2.2 min
- 10Claude Sonnet 5.5 (low)2.5 min
- 11Muse Ember (xhigh)2.7 min
- 12North Mini Code2.8 min
- 13Claude Opus 5.5 (low)3 min
- 14Command A+3 min
- 15Claude Haiku 4.5 (thinking)4 min
| # | ||||||
|---|---|---|---|---|---|---|
| 1 |  Solar Pro 4 | 62.9 min | 48.9 | 311k | $0.956 | AA |
| 2 |  MiMo-V2.6-Flashopen ↗ | 52.9 min | 54.8 | 132k | $0.129 | AA |
| 3 |  Qwen3.8 Maxmax | 51.2 min | 58.8 | 171k | $10.02 | AA |
| 4 |  GLM 5.3 Flashopen ↗ | 48.5 min | 56.5 | 134k | $0.583 | AA |
| 5 |  Ling 3.1 Flashopen ↗ | 46.7 min | 50.8 | 225k | $3.19 | derived |
| 6 |  Kimi K3.1 Proopen ↗ | 39 min | 60.9 | 72k | $3.68 | AA |
| 7 | MiMo-V2.6-Proopen ↗ | 35.9 min | 56.8 | 121k | $0.257 | AA |
| 8 | Qwen3.8-Flash-Nextopen ↗ | 35.8 min | 53.9 | 148k | $0.569 | AA |
| 9 |  Gemini 4 Argonhigh | 30.1 min | 65.3 | 116k | $4.97 | AA |
| 10 | New Inkling 2 Small | 27.7 min | 57.9 | 150k | $1.18 | AA |
| 11 | New Agnes 2.5 Pro Theta | 26 min | 50.8 | 278k | $0.575 | AA |
| 12 | Gemini 4 Carbonlow | 24.9 min | 60.1 | 86k | $5.22 | AA |
| 13 |  Claude Opus 5max | 24.1 min | 64.9 | 118k | $11.82 | AA |
| 14 | Qwen3.8 27Bno reasoningopen ↗ | 19.4 min | 45.5 | 45k | $3.77 | AA |
| 15 |  MiniMax M3.1 Flashopen ↗ | 16.3 min | 53.5 | 138k | $0.491 | AA |
| 16 | Kimi K3lowopen ↗ | 16 min | 53.6 | 40k | $2.37 | AA |
| 17 |  DeepSeek V4 Prono reasoningopen ↗ | 15.9 min | 42.2 | 72k | $0.977 | AA |
| 18 |  Grok 4.6xhigh | 15.3 min | 57.3 | 58k | $4.68 | AA |
| 19 | GLM 5.3lowopen ↗ | 14.6 min | 47.3 | 52k | $1.57 | AA |
| 20 |  Hy4open ↗ | 14.3 min | 56.7 | 78k | $1.03 | AA |
| 21 | Apodex 1.1 | 14.3 min | 48.0 | 116k | $0.764 | AA |
| 22 | MiniMax M3open ↗ | 14.2 min | 56.1 | 91k | $1.29 | AA |
| 23 |  Step 5 Previewopen | 14 min | 54.3 | 113k | $2.35 | AA |
| 24 | Gemini 3.8 Flashmedium | 12.9 min | 60.2 | 82k | $6.26 | AA |
| 25 |  GPT-5.6 Terramax | 11.4 min | 60.4 | 66k | $3.73 | AA |
| 26 | Grok 4.7low | 11.1 min | 55.4 | 43k | $2.92 | AA |
| 27 |  Muse Spark 1.3max | 11 min | 61.5 | 91k | $2.39 | AA |
| 28 | Muse Wonderhigh | 10.9 min | 56.2 | 68k | $4.56 | AA |
| 29 | Inklingxhighopen ↗ | 10.2 min | 50.0 | 51k | $1.35 | AA |
| 30 |  Granite 4.2 8Bopen ↗ | 9.9 min | 41.1 | 52k | $0.040 | AA |
| 31 |  Mistral Large 4mediumopen ↗ | 9.7 min | – | 56k | $1.41 | AA |
| 32 | GPT-5.6 Lunamax | 9.1 min | 56.2 | 59k | $0.331 | AA |
| 33 | gpt-oss-120bhighopen ↗ | 7.9 min | 44.9 | 73k | $0.345 | AA |
| 34 | GPT-6 Solmax | 7.8 min | 62.7 | 54k | $2.47 | AA |
| 35 | Muse Spark Openxhighopen ↗ | 7.8 min | 53.5 | 66k | $2.02 | AA |
| 36 | DeepSeek V4.1 Flash 1017no reasoningopen ↗ | 7.6 min | 44.9 | 65k | $0.425 | AA |
| 37 | Claude Fable 5.1low | 7.5 min | 58.8 | 38k | $4.47 | AA |
| 38 | Inkling Smallopen ↗ | 6.4 min | 52.0 | 37k | $0.203 | AA |
| 39 | NewClaude Fable 5.2low | 6 min | 65.3 | 36k | $3.10 | AA |
| 40 |  Nemotron 3.5 Super VLopen ↗ | 5.7 min | 45.8 | 81k | $0.520 | AA |
| 41 | Muse Glimmerhighopen ↗ | 5.4 min | 46.5 | 26k | $0.170 | AA |
| 42 |  K2 Horizon 375B A23Bopen | 4.7 min | 49.5 | 64k | – | AA |
| 43 | Mistral Medium 3.5open ↗ | 4.6 min | 46.6 | 40k | $1.04 | AA |
| 44 | Nemotron 3 Ultra 550B A55Bthinkingopen ↗ | 4.4 min | 49.1 | 51k | $2.23 | AA |
| 45 | Nemotron 3.5 Lightningopen ↗ | 4.3 min | 41.4 | 65k | $0.314 | AA |
| 46 | GPT-6 Astralow | 4 min | 62.0 | 9k | $1.42 | AA |
| 47 | Claude Haiku 4.5thinking | 4 min | 45.3 | 25k | $0.532 | AA |
| 48 |  Command A+open ↗ | 3 min | 45.9 | 36k | – | AA |
| 49 | Claude Opus 5.5low | 3 min | 58.9 | 19k | $0.976 | AA |
| 50 | North Mini Codeopen | 2.8 min | 40.2 | 30k | – | AA |
| 51 | NewMuse Emberxhigh | 2.7 min | 49.1 | 72k | $0.116 | AA |
| 52 | Claude Sonnet 5.5low | 2.5 min | 50.7 | 21k | $0.511 | AA |
| 53 | Claude Haiku 5.5low | 2.2 min | 47.7 | 28k | $0.044 | AA |
| 54 |  Beam 1.0open | 2.1 min | 46.0 | 34k | $0.284 | AA |
| 55 | GPT-6.1 Sollow | 2 min | 60.0 | 6k | $0.161 | AA |
| 56 | Gemini 3.5 Flash Lite | 1.7 min | 51.4 | 30k | $0.693 | AA |
| 57 | Mistral Small 4thinkingopen | 1.7 min | 40.5 | 14k | $0.040 | AA |
| 58 | NewGemini 4 Hydrogenlow | 80 s | – | 14k | $0.292 | AA |
| 59 | GPT-6.1 Terralow | 70 s | 56.6 | 4k | $0.071 | AA |
| 60 | Gemini 3.1 Flash Lite | 53 s | 52.4 | 16k | $0.194 | AA |
Cite as: BenchLeader, “Time per task: how long one benchmark task takes”, data as of 13 Nov 2026, https://www.benchleader.com/leaderboards/time-per-task.