ProgramBench (Vals AI) leaderboard
- Claude Opus 5.5 leads
- 18.5%
- models measured
- 59
- lead over #2
- +11.5 pts
- median
- 0.0%
Leader on ProgramBench (Vals AI) · read October 21
# ProgramBench (Vals AI), read October 21
leaders = {
"Claude Opus 5.5": 18.5%, "Claude Fable 5.1": 7.0%, "Claude Sonnet 5.5": 6.5%,
}
On October 21, 2026, 59 of our catalog's models had a ProgramBench (Vals AI) result, and Anthropic Claude Opus 5.5 led with 18.5%, ahead of Anthropic Claude Fable 5.1 at 7.0%.
Vals AI's ProgramBench board, run with one shared harness; its page describes the tasks. The score as Vals AI publishes it.
Role in the index
A coding board we show beside the Vibe Coding Index, not inside it: the index keeps the six boards it was defined on, so a score does not move every time a board is added. Read the methodology
| # | Model | Result | Cost per test | Duration, as printed | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Opus 5.5Anthropic | 18.5% | $68.05 | 2h 24m | 88.8 | $20.00 | |
| 2 | Claude Fable 5.1Anthropic | 7.0% | $58.52 | 2h 27m | 84.2 | $50.00 | |
| 3 | Claude Sonnet 5.5Anthropic | 6.5% | $35.86 | 2h 07m | 87.3 | $10.00 | |
| 4 | GPT-6 AstraOpenAI | 5.5% | $11.57 | 27m 02s | 85.1 | $50.00 | |
| 5 | Claude Opus 5Anthropic | 3.0% | $60.29 | 2h 55m | 83.0 | $25.00 | |
| 6 | GPT-6.1 SolOpenAI | 3.0% | $2.87 | 1h 10m | 82.4 | $10.00 | |
| 7 | Claude Haiku 5.5Anthropic | 2.5% | $6.83 | 1h 13m | 70.7 | $0.50 | |
| 8 | Gemini 4 ArgonGoogle | 2.5% | $17.66 | 1h 11m | 81.0 | $10.00 | |
| 9 | Claude Fable 5Anthropic | 2.0% | $75.68 | 2h 37m | 79.3 | $50.00 | |
| 10 | Kimi K3Moonshot AI | 2.0% | $70.48 | 5h 42m | 62.3 | $15.00 | |
| 11 | GPT-6 SolOpenAI | 2.0% | $11.67 | 46m 11s | 76.0 | $10.00 | |
| 12 | GPT-5.6 SolOpenAI | 1.5% | $15.29 | 29m 24s | 73.5 | $20.00 | |
| 13 | GLM 5.3Z.ai | 1.5% | $21.96 | 4h 16m | 70.6 | $4.40 | |
| 14 | Claude Opus 4.8Anthropic | 1.0% | $31.27 | 1h 32m | 65.0 | $25.00 | |
| 15 | Gemini 3.8 FlashGoogle | 1.0% | $10.93 | 46m 39s | 61.6 | $3.75 | |
| 16 | Grok 4.6xAI | 1.0% | $16.76 | 4h 27m | 65.7 | $6.00 | |
| 17 | Claude Sonnet 4.6Anthropic | 0.5% | Not published | 1h 12m | 57.3 | $15.00 | |
| 18 | DeepSeek V4.1 FlashDeepSeek | 0.5% | $0.90 | 1h 02m | 62.5 | $0.78 | |
| 19 | Muse Spark 1.2Meta | 0.5% | $2.47 | 26m 38s | 55.7 | $4.25 | |
| 20 | Muse Spark 1.3Meta | 0.5% | $11.56 | 1h 09m | 57.4 | $4.25 | |
| 21 | GPT-5.4OpenAI | 0.5% | $2.27 | 15m 58s | 64.4 | $15.00 | |
| 22 | GPT-5.5OpenAI | 0.5% | $6.95 | 22m 03s | 71.3 | $30.00 | |
| 23 | GPT-5.6 TerraOpenAI | 0.5% | $5.38 | 33m 36s | 63.9 | $12.00 | |
| 24 | GPT-6 LunaOpenAI | 0.5% | $0.18 | 28m 55s | 57.3 | $0.50 | |
| 25 | Grok 4.7xAI | 0.5% | $46.49 | 3h 38m | 66.7 | $6.00 | |
| 26 | T Inkling SmallThinking Machines | 0.5% | $0.48 | 28m 31s | 27.3 | $1.20 | |
| 27 | MiMo-V2.6-FlashXiaomi | 0.5% | $0.47 | 3h 22m | 54.7 | $0.28 | |
| 28 | MiMo-V2.6-ProXiaomi | 0.5% | $0.73 | 3h 00m | 65.4 | $0.87 | |
| 29 | GLM 5.2Z.ai | 0.5% | $12.58 | 2h 06m | 61.1 | $4.40 | |
| 30 | Claude Haiku 4.5Anthropic | 0.0% | Not published | 7m 37s | 35.6 | $5.00 | |
| 31 | Claude Opus 4.7Anthropic | 0.0% | Not published | 54m 07s | 67.5 | $25.00 | |
| 32 | Claude Sonnet 5Anthropic | 0.0% | $24.36 | 1h 31m | 57.7 | $10.00 | |
| 33 | DeepSeek V4 Flash 0731DeepSeek | 0.0% | $1.29 | 5h 39m | 51.8 | $0.28 | |
| 34 | DeepSeek V4 Pro 0423DeepSeek | 0.0% | Not published | 1h 03m | 38.7 | $2.716 | |
| 35 | DeepSeek V4 Pro 0813DeepSeek | 0.0% | $0.43 | 1h 23m | 56.1 | $1.98 | |
| 36 | Gemini 3 Flash PreviewGoogle | 0.0% | $0.44 | 6m 59s | 43.7 | $3.00 | |
| 37 | Gemini 3.1 Flash Lite PreviewGoogle | 0.0% | $0.06 | 2m 56s | 22.7 | $1.50 | |
| 38 | Gemini 3.1 Pro PreviewGoogle | 0.0% | $2.63 | 12m 36s | 32.3 | $12.00 | |
| 39 | Gemini 3.5 FlashGoogle | 0.0% | Not published | 18m 25s | 39.8 | $9.00 | |
| 40 | Gemini 3.5 Flash LiteGoogle | 0.0% | $0.34 | 7m 15s | 49.0 | $2.50 | |
| 41 | Gemini 3.6 FlashGoogle | 0.0% | $5.92 | 47m 33s | 62.5 | $3.75 | |
| 42 | Gemini 3.7 FlashGoogle | 0.0% | $7.13 | 21m 30s | 58.0 | $3.75 | |
| 43 | Muse Spark 1.1Meta | 0.0% | $13.06 | 5h 43m | 70.3 | $4.25 | |
| 44 | MiniMax M2.7MiniMax | 0.0% | Not published | 24m 20s | 36.2 | $1.20 | |
| 45 | Kimi K2.6Moonshot AI | 0.0% | Not published | 1h 20m | 52.1 | $4.00 | |
| 46 | Kimi K2.7 CodeMoonshot AI | 0.0% | $3.19 | 1h 04m | 53.4 | $4.00 | |
| 47 | Nemotron 3 UltraNVIDIA | 0.0% | Not published | 1h 02m | 30.0 | $2.40 | |
| 48 | Nemotron 3.5 LightningNVIDIA | 0.0% | $0.22 | 24m 26s | 22.6 | $0.147 | |
| 49 | GPT-5.4 MiniOpenAI | 0.0% | Not published | 53m 45s | 35.8 | $4.50 | |
| 50 | GPT-5.6 LunaOpenAI | 0.0% | $0.94 | 41m 34s | 58.0 | $1.20 | |
| 51 | Qwen3.6 PlusQwen | 0.0% | Not published | 33m 01s | 44.1 | $1.95 | |
| 52 | Qwen3.8 27BQwen | 0.0% | $17.99 | 5h 57m | 74.2 | $2.55 | |
| 53 | Qwen3.8 Max (0902)Qwen | 0.0% | $12.36 | 5h 54m | 62.4 | $6.00 | |
| 54 | Grok 4.3xAI | 0.0% | $0.36 | 3m 56s | 37.4 | $2.50 | |
| 55 | Grok 4.5xAI | 0.0% | $4.12 | 18m 59s | 58.6 | $6.00 | |
| 56 | Hy4 previewTencent | 0.0% | $12.81 | 2h 10m | 44.3 | $2.2509 | |
| 57 | T InklingThinking Machines | 0.0% | $4.14 | 45m 34s | 31.5 | $4.05 | |
| 58 | GLM 5.1Z.ai | 0.0% | Not published | 1h 06m | 49.2 | $4.40 | |
| 59 | GLM 5.3 FlashZ.ai | 0.0% | $0.80 | 5h 30m | 57.7 | $0.50 |
Only models in our catalog are listed, and each appears once with its best published run. 136 rows were read from the source on October 21, 2026, and rows that match no catalog model are left out. Cost per test is what Vals AI reports one test cost on the run behind the result, in US dollars. Every other column is a figure the board publishes for the same run.