LiveCodeBench leaderboard
As of 15 Nov 2026, the top score on LiveCodeBench is 90.5% by Claude Fable 5.1 (Anthropic), ahead of Claude Fable 5 at 89.8%, across 140 model configurations with a published result (Vals AI, result dated 1 Sep 2026). LiveCodeBench collects competitive-programming problems published after each model's training cutoff and checks solutions by running hidden tests. LiveCodeBench counts toward the BenchLeader coding score.
Key facts (as of 15 Nov 2026)
- Claude Fable 5.1 holds the top LiveCodeBench score at 90.5% (Vals AI, 1 Sep 2026).
- Claude Fable 5 is second on LiveCodeBench at 89.8% (Vals AI, 1 Sep 2026).
- LiveCodeBench: Competitive programming problems released after model cutoffs. Run by Vals AI (Vals AI).
- 140 model configurations have a published LiveCodeBench result on BenchLeader (BenchLeader, 15 Nov 2026).
- LiveCodeBench feeds the coding category of the BenchLeader Index (BenchLeader methodology).
Competitive programming problems released after model cutoffs. Run by Vals AI.
- Published by
- Vals AI
- Category
- Coding
- Index weight
- 1.0
- Models
- 140
- Data as of
- 15 Nov 2026
Vals AI (vals.ai).
What the test looks like
LiveCodeBench collects competitive-programming problems published after each model's training cutoff and checks solutions by running hidden tests.
How it is scored
Percent of problems passing all tests; Vals AI runs a fixed recent window.
What to keep in mind
Only the window after a model's cutoff is contamination-free, and the window changes over time, so scores across dates are not strictly comparable.
- 1Claude Fable 5.1 (max)90.5%
- 2Claude Fable 5 (max)89.8%
- 3Gemini 3.8 Flash (high)89.5%
- 4Claude Opus 5 (max)89.0%
- 5Gemini 3.7 Flash (high)88.7%
- 6Gemini 3.1 Pro (high)88.5%
- 7Grok 4.6 (high)88.2%
- 8Gemini 3.6 Flash (high)88.1%
- 9GPT-5.2-Codex88.0%
- 10Qwen3.8 Max (max)87.8%
- 11Claude Opus 4.8 (max)87.8%
- 12Gemini 3.5 Flash (high)87.6%
- 13DeepSeek V4 Pro (max)87.5%
- 14Grok 4.5 (high)87.3%
- 15GPT-5.3-Codex (xhigh)87.3%
| # | |||||
|---|---|---|---|---|---|
| 1 | 90.5% | 66.7 | $0.261 | 2026-09-01 | |
| 2 | 89.8% | 65.1 | $0.428 | 2026-09-01 | |
| 3 | 89.5% | 59.4 | $0.071 | 2026-09-01 | |
| 4 | 89.0% | 65.0 | $0.134 | 2026-09-01 | |
| 5 | 88.7% | 62.1 | $0.040 | 2026-09-01 | |
| 6 | 88.5% | 60.4 | $0.101 | 2026-09-01 | |
| 7 | 88.2% | 60.6 | $0.040 | 2026-09-01 | |
| 8 | 88.1% | 61.0 | $0.056 | 2026-09-01 | |
| 9 | 88.0% | 57.6 | $0.117 | 2026-09-01 | |
| 10 | 87.8% | 58.7 | $0.091 | 2026-09-01 | |
| 11 | 87.8% | 61.1 | $0.327 | 2026-09-01 | |
| 12 | 87.6% | 62.5 | $0.086 | 2026-09-01 | |
| 13 | 87.5% | 54.7 | $0.064 | 2026-09-01 | |
| 14 | 87.3% | 56.8 | $0.045 | 2026-09-01 | |
| 15 | 87.3% | 62.9 | $0.092 | 2026-09-01 | |
| 16 | 87.3% | 57.2 | $0.020 | 2026-09-01 | |
| 17 | 87.2% | 60.2 | $0.104 | 2026-09-01 | |
| 18 | 87.1% | 60.0 | $0.087 | 2026-09-01 | |
| 19 | 86.8% | 55.1 | $0.065 | 2026-09-01 | |
| 20 | 86.6% | 54.6 | $0.012 | 2026-09-01 | |
| 21 | 86.5% | 57.5 | $0.126 | 2026-09-01 | |
| 22 | 86.4% | 60.0 | $0.127 | 2026-09-01 | |
| 23 | 86.0% | 55.4 | – | 2026-09-01 | |
| 24 | 86.0% | 59.0 | $0.078 | 2026-09-01 | |
| 25 | 85.9% | 60.1 | $0.031 | 2026-09-01 | |
| 26 | 85.9% | 52.0 | $0.021 | 2026-09-01 | |
| 27 | 85.9% | 59.5 | $0.046 | 2026-09-01 | |
| 28 | 85.9% | 58.2 | $0.064 | 2026-09-01 | |
| 29 | 85.6% | 57.4 | $0.051 | 2026-09-01 | |
| 30 | 85.5% | – | $0.138 | 2026-09-01 | |
| 31 | 85.5% | 50.4 | $0.153 | 2026-09-01 | |
| 32 | 85.4% | 59.9 | $0.135 | 2026-09-01 | |
| 33 | 85.3% | 57.4 | $0.048 | 2026-09-01 | |
| 34 | 85.3% | 64.1 | $0.067 | 2026-09-01 | |
| 35 | 85.1% | 59.7 | $0.351 | 2026-09-01 | |
| 36 | 84.7% | 58.6 | $0.097 | 2026-09-01 | |
| 37 | 84.7% | 57.0 | $0.343 | 2026-09-01 | |
| 38 | 84.5% | 56.5 | $0.020 | 2026-09-01 | |
| 39 | 84.3% | 58.3 | $0.052 | 2026-09-01 | |
| 40 | 84.1% | 63.1 | $0.329 | 2026-09-01 | |
| 41 | 84.0% | 48.3 | $0.0025 | 2026-09-01 | |
| 42 | 84.0% | 52.2 | $0.045 | 2026-09-01 | |
| 43 | 84.0% | 51.7 | $0.0030 | 2026-09-01 | |
| 44 | 83.9% | 55.1 | $0.033 | 2026-09-01 | |
| 45 | 83.9% | 59.4 | $0.039 | 2026-09-01 | |
| 46 | 83.7% | 59.7 | $0.971 | 2026-09-01 | |
| 47 | 83.6% | – | – | 2026-09-01 | |
| 48 | 83.3% | 51.8 | $0.0095 | 2026-09-01 | |
| 49 | 83.3% | 58.1 | $0.156 | 2026-09-01 | |
| 50 | 83.2% | 46.1 | $0.0085 | 2026-09-01 | |
| 51 | 82.6% | 65.3 | $0.090 | 2026-09-01 | |
| 52 | 82.4% | 56.1 | $0.095 | 2026-09-01 | |
| 53 | 82.2% | 51.7 | $0.035 | 2026-09-01 | |
| 54 | 82.2% | 51.2 | $0.014 | 2026-09-01 | |
| 55 | 82.2% | 56.1 | $0.037 | 2026-09-01 | |
| 56 | 82.1% | 56.1 | $0.371 | 2026-09-01 | |
| 57 | 82.0% | 53.9 | $0.027 | 2026-09-01 | |
| 58 | 81.9% | 60.2 | $0.044 | 2026-09-01 | |
| 59 | 81.8% | 53.8 | $0.018 | 2026-09-01 | |
| 60 | 81.5% | 56.3 | $0.0022 | 2026-09-01 |
AI coding benchmark leaderboard · All LLM benchmarks
Related benchmarks
Top head-to-heads on LiveCodeBench
Cite as: BenchLeader, “LiveCodeBench leaderboard”, data as of 15 Nov 2026, https://www.benchleader.com/benchmarks/vals_livecodebench.
LiveCodeBench: questions
- What does LiveCodeBench measure?
- LiveCodeBench collects competitive-programming problems published after each model's training cutoff and checks solutions by running hidden tests. Scores are reported in percent of tasks solved; higher is better.
- What is the top score on LiveCodeBench?
- As of 15 Nov 2026, the highest score on LiveCodeBench is 90.5% by Claude Fable 5.1 (result dated 1 Sep 2026), ahead of Claude Fable 5 at 89.8%. Scores come from Vals AI and are re-read daily.
- Which AI model leads LiveCodeBench?
- Claude Fable 5.1 leads LiveCodeBench with 90.5% as of 15 Nov 2026 (result dated 1 Sep 2026), ahead of Claude Fable 5 at 89.8%.
- How many models have LiveCodeBench results?
- 140 model configurations have a LiveCodeBench result on BenchLeader, all taken from Vals AI.
- Who runs LiveCodeBench and how often is it updated?
- LiveCodeBench is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
- Does LiveCodeBench count toward the BenchLeader Index?
- Yes. LiveCodeBench contributes to the coding category of the BenchLeader Index, normalised so that 50 is the average of the evaluated models.