Vibe Code Bench (Vals AI) leaderboard
- Claude Sonnet 5.5 leads
- 92.4%
- models measured
- 96
- lead over #2
- +0.5 pts
- median
- 54.6%
Leader on Vibe Code Bench (Vals AI) · read October 22
# Vibe Code Bench (Vals AI), read October 22
leaders = {
"Claude Sonnet 5.5": 92.4%, "Gemini 4 Argon": 91.9%, "Claude Haiku 5.5": 90.4%,
}
On October 22, 2026, 96 of our catalog's models had a Vibe Code Bench (Vals AI) result, and Anthropic Claude Sonnet 5.5 led with 92.4%, ahead of Google Gemini 4 Argon at 91.9%.
Vals AI's Vibe Code board, run with one shared harness; its page describes the tasks. The score as Vals AI publishes it.
Role in the index
Vibe Code Bench (Vals AI) carries 20% of the Vibe Coding Index. A model missing it is scored on the other benchmarks of that index. Read the methodology
| # | Model | Result | Cost per test | Duration, as printed | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Claude Sonnet 5.5Anthropic | 92.4% | $31.25 | 1h 01m | 87.3 | $10.00 | |
| 2 | Gemini 4 ArgonGoogle | 91.9% | $8.11 | 37m 28s | 81.0 | $10.00 | |
| 3 | Claude Haiku 5.5Anthropic | 90.4% | $6.07 | 34m 36s | 70.7 | $0.50 | |
| 4 | Claude Fable 5Anthropic | 90.3% | $41.71 | 1h 02m | 79.3 | $50.00 | |
| 5 | Claude Opus 5.5Anthropic | 90.3% | $57.92 | 1h 36m | 88.8 | $20.00 | |
| 6 | Claude Fable 5.1Anthropic | 90.3% | $33.37 | 57m 40s | 84.2 | $50.00 | |
| 7 | GPT-6 AstraOpenAI | 89.6% | $38.51 | 39m 04s | 85.1 | $50.00 | |
| 8 | GPT-6.1 SolOpenAI | 88.9% | $6.23 | 1h 15m | 82.4 | $10.00 | |
| 9 | Claude Opus 5Anthropic | 88.4% | $33.88 | 1h 28m | 83.0 | $25.00 | |
| 10 | GPT-6 SolOpenAI | 87.8% | $26.36 | 38m 56s | 76.0 | $10.00 | |
| 11 | DeepSeek V4.1 Flash 1017DeepSeek | 87.7% | $0.58 | 16m 14s | 70.3 | $0.60 | |
| 12 | Grok 4.7xAI | 86.2% | $15.83 | 35m 47s | 66.7 | $6.00 | |
| 13 | MiMo-V2.6-ProXiaomi | 85.2% | $1.04 | 1h 01m | 65.4 | $0.87 | |
| 14 | Kimi K3Moonshot AI | 85.0% | $10.01 | 16m 39s | 62.3 | $15.00 | |
| 15 | DeepSeek V4.1 FlashDeepSeek | 84.7% | $0.41 | 15m 28s | 62.5 | $0.84 | |
| 16 | Ember-1Fireworks | 83.9% | $8.56 | 39m 18s | 52.6 | $15.00 | |
| 17 | Muse WonderMeta | 83.1% | $1.80 | 5m 50s | 71.9 | $20.00 | |
| 18 | Muse Spark 1.3Meta | 82.9% | $2.10 | 12m 54s | 57.4 | $4.25 | |
| 19 | Claude Opus 4.8Anthropic | 82.7% | $26.88 | 1h 16m | 65.0 | $25.00 | |
| 20 | DeepSeek V4 Pro 0813DeepSeek | 82.3% | $0.36 | 1h 06m | 56.1 | $1.98 | |
| 21 | GPT-6 LunaOpenAI | 81.7% | $1.35 | 37m 00s | 57.3 | $0.50 | |
| 22 | Claude Sonnet 5Anthropic | 81.3% | $25.39 | 1h 07m | 57.7 | $10.00 | |
| 23 | GPT-5.6 SolOpenAI | 80.5% | $33.40 | 33m 53s | 73.5 | $20.00 | |
| 24 | Muse Spark 1.2Meta | 79.1% | $1.53 | 19m 55s | 55.7 | $4.25 | |
| 25 | MiMo-V2.6-FlashXiaomi | 79.0% | $0.56 | 50m 13s | 54.7 | $0.28 | |
| 26 | Gemini 3.8 FlashGoogle | 78.7% | $6.87 | 8m 39s | 61.6 | $3.75 | |
| 27 | Mistral Large 4Mistral AI | 78.4% | $39.46 | 2h 53m | 56.4 | $4.18 | |
| 28 | GLM 5.3Z.ai | 78.1% | $12.45 | 1h 04m | 70.6 | $4.40 | |
| 29 | Hy4 previewTencent | 77.5% | $2.18 | 41m 33s | 44.3 | $2.501 | |
| 30 | GPT-5.6 LunaOpenAI | 77.1% | $0.73 | 25m 52s | 58.0 | $1.20 | |
| 31 | Grok 4.6xAI | 76.2% | $4.88 | 25m 30s | 65.7 | $6.00 | |
| 32 | Muse Spark OpenMeta | 76.2% | $1.66 | 31m 09s | 47.4 | $4.25 | |
| 33 | DeepSeek V4 Flash 0731DeepSeek | 74.7% | $0.91 | 37m 22s | 51.8 | $0.396 | |
| 34 | GPT-5.6 TerraOpenAI | 74.6% | $7.89 | 19m 12s | 63.9 | $12.00 | |
| 35 | Muse Spark 1.1Meta | 72.2% | $0.92 | 10m 15s | 70.3 | $4.25 | |
| 36 | Claude Opus 4.7Anthropic | 71.0% | $21.41 | 35m 42s | 67.5 | $25.00 | |
| 37 | Step 5 PreviewStepFun | 70.4% | $5.86 | 1h 31m | 61.1 | $2.70 | |
| 38 | Gemini 3.7 FlashGoogle | 70.4% | $4.83 | 23m 43s | 58.0 | $3.75 | |
| 39 | GPT-5.5OpenAI | 69.8% | $16.66 | 31m 52s | 71.3 | $30.00 | |
| 40 | Grok 4.5xAI | 69.0% | $4.10 | 13m 50s | 58.6 | $6.00 | |
| 41 | GPT-5.4OpenAI | 67.4% | $10.69 | 1h 26m | 64.4 | $15.00 | |
| 42 | Qwen3.8 27BQwen | 64.8% | $23.46 | 43m 39s | 74.2 | $2.55 | |
| 43 | Qwen3.8 Max (0902)Qwen | 64.7% | $8.24 | 2h 49m | 62.4 | $6.00 | |
| 44 | Gemini 3.6 FlashGoogle | 64.0% | $3.04 | 25m 23s | 62.5 | $3.75 | |
| 45 | GLM 5.2Z.ai | 64.0% | $8.49 | 1h 02m | 61.1 | $4.40 | |
| 46 | GPT-5.3-CodexOpenAI | 61.8% | $11.91 | 1h 16m | 60.0 | $14.00 | |
| 47 | Claude Opus 4.6Anthropic | 57.6% | $8.69 | 21m 19s | 57.2 | $25.00 | |
| 48 | Claude Sonnet 4.6Anthropic | 55.8% | $7.35 | 49m 42s | 57.3 | $15.00 | |
| 49 | GPT-5.2OpenAI | 53.5% | $17.75 | 1h 23m | 55.9 | $14.00 | |
| 50 | DeepSeek V4 Pro 0423DeepSeek | 49.9% | $2.21 | 56m 59s | 38.7 | $2.976 | |
| 51 | Gemini 3.5 FlashGoogle | 48.7% | $2.54 | 14m 53s | 39.8 | $9.00 | |
| 52 | GPT-5.4 MiniOpenAI | 48.0% | $1.19 | 34m 15s | 35.8 | $4.50 | |
| 53 | Qwen3.7 MaxQwen | 47.7% | $11.40 | 1h 18m | 54.1 | $4.425 | |
| 54 | MiniMax M3MiniMax | 47.6% | $6.45 | 1h 25m | 39.2 | $1.20 | |
| 55 | Kimi K2.7 CodeMoonshot AI | 47.2% | $3.84 | 2h 52m | 53.4 | $4.00 | |
| 56 | Qwen3.7 PlusQwen | 46.4% | $1.08 | 37m 12s | 33.2 | $1.28 | |
| 57 | MiMo-V2.5Xiaomi | 42.2% | $0.07 | 26m 20s | 50.0 | $0.336 | |
| 58 | GPT-5.2-CodexOpenAI | 37.9% | $4.15 | 32m 15s | 45.4 | $14.00 | |
| 59 | Kimi K2.6Moonshot AI | 37.9% | $1.93 | 49m 28s | 52.1 | $4.00 | |
| 60 | Gemini 3.5 Flash LiteGoogle | 37.2% | $0.83 | 8m 22s | 49.0 | $2.50 | |
| 61 | MiMo-V2.5-ProXiaomi | 34.1% | $0.17 | 40m 04s | 31.6 | $0.9152 | |
| 62 | Gemini 3.1 Pro PreviewGoogle | 32.0% | $3.83 | 20m 12s | 32.3 | $12.00 | |
| 63 | GLM 5.1Z.ai | 31.5% | $2.89 | 33m 35s | 49.2 | $4.40 | |
| 64 | GLM 5.3 FlashZ.ai | 30.8% | $2.35 | 1h 24m | 57.7 | $0.50 | |
| 65 | GPT-5.4 NanoOpenAI | 26.1% | $1.28 | 54m 36s | 44.8 | $1.25 | |
| 66 | Qwen3.6 PlusQwen | 25.6% | $5.45 | 38m 10s | 44.1 | $1.95 | |
| 67 | GPT-5.1OpenAI | 24.6% | $2.57 | 30m 36s | 40.4 | $10.00 | |
| 68 | GLM 5Z.ai | 23.4% | $40.27 | 3h 44m | 41.7 | $3.20 | |
| 69 | Claude Sonnet 4.5Anthropic | 22.6% | $6.66 | 49m 22s | 39.5 | $15.00 | |
| 70 | GPT-5.1-Codex-MaxOpenAI | 22.2% | $7.32 | 8h 40m | Not benchmarked | $10.00 | |
| 71 | Claude Opus 4.5Anthropic | 20.6% | $32.87 | 38m 03s | 44.0 | $25.00 | |
| 72 | Gemini 3 Flash PreviewGoogle | 20.2% | $0.94 | 13m 27s | 43.7 | $3.00 | |
| 73 | GPT-5OpenAI | 20.1% | $1.53 | 30m 52s | 38.8 | $10.00 | |
| 74 | Grok 4.3xAI | 19.4% | $1.28 | 9m 49s | 37.4 | $2.50 | |
| 75 | T InklingThinking Machines | 19.2% | $1.61 | 18m 25s | 31.5 | $4.05 | |
| 76 | T Inkling SmallThinking Machines | 19.1% | $0.43 | 49m 09s | 27.3 | $1.20 | |
| 77 | Kimi K2.5Moonshot AI | 17.5% | $0.88 | 42m 50s | 38.6 | $2.85 | |
| 78 | Qwen3.5 Plus 2026-04-20Qwen | 15.7% | $3.80 | 50m 16s | 39.5 | $1.80 | |
| 79 | MiniMax M2.5MiniMax | 14.8% | $2.20 | 51m 06s | 37.2 | $1.20 | |
| 80 | GPT-5 MiniOpenAI | 14.2% | $0.25 | 11m 38s | 34.2 | $2.00 | |
| 81 | Grok Build 0.1xAI | 13.4% | Not published | 6m 35s | Not benchmarked | $2.00 | |
| 82 | GPT-5.1-CodexOpenAI | 13.1% | $3.80 | 50m 26s | 15.3 | $10.00 | |
| 83 | Qwen3.6 27BQwen | 11.9% | $4.66 | 2h 43m | 36.8 | $2.70 | |
| 84 | MiniMax M2.7MiniMax | 11.9% | $2.82 | 22m 57s | 36.2 | $1.20 | |
| 85 | Claude Haiku 4.5Anthropic | 11.4% | $1.31 | 12m 56s | 35.6 | $5.00 | |
| 86 | Nemotron 3 UltraNVIDIA | 7.6% | Not published | 36m 57s | 30.0 | $2.40 | |
| 87 | DeepSeek V3.2DeepSeek | 5.1% | $2.47 | 56m 06s | 30.0 | $0.88 | |
| 88 | Grok 4.20xAI | 4.1% | $0.77 | 5m 02s | 33.3 | $2.50 | |
| 89 | Mercury 2.5Inception | 3.7% | $0.38 | 3m 13s | 1.6 | $0.15 | |
| 90 | GLM 4.6Z.ai | 3.1% | $10.85 | 2h 47m | 8.8 | $2.20 | |
| 91 | Ling 3.0 FlashinclusionAI | 2.9% | $0.29 | 16m 46s | 29.6 | $0.1215 | |
| 92 | Mistral Medium 3.5Mistral AI | 2.9% | $16.74 | 41m 11s | 25.5 | $7.50 | |
| 93 | Gemini 2.5 ProGoogle | 0.4% | $1.22 | 34m 57s | 19.1 | $10.00 | |
| 94 | Gemini 3.1 Flash Lite PreviewGoogle | 0.0% | $0.51 | 9m 33s | 22.7 | $1.50 | |
| 95 | Mistral Small 4Mistral AI | 0.0% | $6.47 | 30m 15s | Not benchmarked | $0.60 | |
| 96 | Nemotron 3.5 LightningNVIDIA | 0.0% | $0.13 | 16m 00s | 22.6 | $0.147 |
Only models in our catalog are listed, and each appears once with its best published run. 137 rows were read from the source on October 22, 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.