Skip to main content
BenchLM
Advertise

Benchmark profile

Vals-hosted LiveCodeBench mirror (Vals LiveCodeBench mirror)

Vals AI implementation of LiveCodeBench with easy, medium, and hard task splits.

Data verified

How BenchLM shows Vals LiveCodeBench mirror

BenchLM mirrors the public Vals AI Vals LiveCodeBench mirror leaderboard captured from https://www.vals.ai/benchmarks/lcb and updated by Vals on July 23, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

Vals LiveCodeBench mirror is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

131 Vals rows4 task viewspublic datasetTasks: Overall, Easy, Medium, HardDisplay only
Vals LiveCodeBench mirror on Vals AIVals methodologyVals home

Vals LiveCodeBench mirror score on Vals LiveCodeBench mirror — July 23, 2026

BenchLM mirrors the published vals livecodebench mirror score view for Vals LiveCodeBench mirror. Claude Fable 5 leads the public snapshot at 89.78% , followed by Claude Opus 5 (89.03%) and Gemini 3.1 Pro Preview (88.48%). BenchLM does not use these results to rank models overall.

1

Claude Fable 5

Anthropic

anthropic/claude-fable-5

89.78%
Overall —
2

Claude Opus 5

Anthropic

anthropic/claude-opus-5

89.03%
Overall —
3

Gemini 3.1 Pro Preview

Google

google/gemini-3.1-pro-preview

88.48%
Overall —
131 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated July 23, 2026

Vals LiveCodeBench mirror score table (131 models)

Score
1
Claude Fable 5Anthropic
89.78%
2
Claude Opus 5Anthropic
89.03%
3
Gemini 3.1 Pro PreviewGoogle
88.48%
4
Gemini 3.6 FlashGoogle
88.08%
5
GPT-5.2 CodexOpenAI
87.99%
6
Claude Opus 4.8Anthropic
87.82%
7
Gemini 3.5 FlashGoogle
87.60%
8
DeepSeek V4 ProDeepSeek
87.48%
9
Grok 4.5xAI
87.35%
10
GPT-5.3 CodexOpenAI
87.31%
11
Kimi K3Moonshot AI
87.19%
12
Qwen3.7 MaxAlibaba
87.06%
13
Kimi K2.6Moonshot AI
86.77%
14
GPT-5 MiniOpenAI
86.61%
15
GPT-5.1OpenAI
86.49%
16
Gemini 3 Pro PreviewGoogle
86.41%
17
Nemotron 3 Ultra 550b A55bNvidia
85.98%
18
Qwen3.6 PlusAlibaba
85.95%
19
GPT-5.6 TerraOpenAI
85.93%
20
GPT-5OpenAI
85.91%
21
Muse Spark 1.1Meta
85.91%
22
Gemini 3 Flash PreviewGoogle
85.59%
23
GPT-5.1 CodexOpenAI
85.55%
24
InklingThinkingmachines
85.52%
25
GPT-5.2OpenAI
85.36%
26
Qwen3.5 Plus ThinkingAlibaba
85.33%
27
GPT-5.5OpenAI
85.30%
28
Claude Opus 4.7Anthropic
85.07%
29
GPT-5 CodexOpenAI
84.72%
30
Claude Opus 4.6 ThinkingAnthropic
84.68%
31
Grok 4.3xAI
84.49%
32
Grok 4.20 0309 ReasoningxAI
84.27%
33
GPT-5.4OpenAI
84.14%
34
GPT-5.4 NanoOpenAI
84.01%
35
O3OpenAI
83.91%
36
Kimi K2.5 ThinkingMoonshot AI
83.87%
37
Claude Opus 4.5 20251101 ThinkingAnthropic
83.67%
38
GPT-5.1 Codex MaxOpenAI
83.56%
39
Qwen3.5 FlashAlibaba
83.28%
40
Grok 4 0709xAI
83.25%
41
GPT Oss 120bFireworks AI
83.23%
42
GPT-5.6 SolOpenAI
82.60%
43
Claude Sonnet 5Anthropic
82.43%
44
GLM 4.7Zhipu AI
82.23%
45
O4 MiniOpenAI
82.21%
46
MiniMax M3MiniMax
82.15%
47
Claude Sonnet 4.6Anthropic
82.09%
48
Kimi K2.7 CodeMoonshot AI
82.05%
49
GLM 5 ThinkingZhipu AI
81.87%
50
MiniMax M2.1MiniMax
81.76%
51
Mimo V2.5Xiaomi
81.51%
52
GPT-5.4 MiniOpenAI
81.47%
53
GLM 5.1Zhipu AI
81.38%
54
Mimo V2.5 ProXiaomi
81.35%
55
GLM 4.6Zhipu AI
81.04%
56
DeepSeek V3p2 ThinkingFireworks AI
80.69%
57
Grok 4.1 Fast ReasoningxAI
80.64%
58
GPT Oss 20bFireworks AI
80.39%
59
Gemini 3.1 Flash Lite PreviewGoogle
80.12%
60
MiniMax M2.7MiniMax
79.93%
61
MiniMax M2.5MiniMax
79.21%
62
Gemini 2.5 Pro Preview 03 25Google
79.16%
63
Gemini 3.5 Flash LiteGoogle
79.01%
64
Grok 4 Fast ReasoningxAI
78.97%
65
Qwen3 MaxAlibaba
78.22%
66
Grok 3 Mini Fast High ReasoningxAI
76.22%
67
Gemini 2.5 Flash Preview 09 2025 ThinkingGoogle
76.21%
68
Gemini 2.5 Flash Preview 09 2025Google
75.06%
69
Claude Opus 4.5Anthropic
75.03%
70
Magistral Medium 2509Mistral AI
74.86%
71
Claude Sonnet 4.5 20250929 ThinkingAnthropic
73.00%
72
Magistral Small 2509Mistral AI
72.13%
73
O3 MiniOpenAI
71.48%
74
Gemini 2.5 Flash Lite Preview 09 2025 ThinkingGoogle
71.39%
75
Qwen3 235b A22bFireworks AI
70.62%
76
Moonshotai Kimi K2 InstructTogether AI
70.45%
77
DeepSeek R1Fireworks AI
70.22%
78
GPT-5 NanoOpenAI
70.22%
79
Claude Opus 4 20250514 ThinkingAnthropic
70.19%
80
DeepSeek V3p2Fireworks AI
69.86%
81
GLM 5.2Zhipu AI
69.50%
82
Gemini 2.5 Flash Lite Preview 09 2025Google
67.67%
83
GLM 4.5Zhipu AI
67.45%
84
Qwen3 Max PreviewAlibaba
66.91%
85
Laguna Xs.2Poolside
66.61%
86
Claude Opus 4.1 20250805 ThinkingAnthropic
66.46%
87
Grok 3 Mini Fast Low ReasoningxAI
66.27%
88
DeepSeek V3 0324Fireworks AI
65.48%
89
Claude Opus 4.1Anthropic
64.56%
90
Kimi K2 ThinkingMoonshot AI
63.15%
91
Claude Opus 4Anthropic
62.63%
92
Claude Sonnet 4 20250514 ThinkingAnthropic
62.39%
93
Grok Code Fast 1xAI
61.97%
94
Laguna M.1Poolside
61.44%
95
Claude 3.7 Sonnet 20250219 ThinkingAnthropic
60.44%
96
Claude Sonnet 4Anthropic
59.67%
97
Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561 ThinkingTogether AI
58.37%
98
GPT-4.1 MiniOpenAI
58.16%
99
Gemini 2.5 Flash Preview 04 17Google
56.94%
100
Claude 3.7 SonnetAnthropic
56.66%
101
Mistral Large 2512Mistral AI
55.34%
102
GPT-4.1OpenAI
54.67%
103
Grok 3xAI
52.90%
104
Devstral 2512Mistral AI
51.84%
105
O1OpenAI
50.26%
106
Claude 3.5 SonnetAnthropic
49.63%
107
Llama4 Maverick Instruct BasicFireworks AI
47.25%
108
Gemini 2.5 Flash Preview 04 17 ThinkingGoogle
46.87%
109
Grok 4 Fast Non ReasoningxAI
46.09%
110
Mistral Medium 2505Mistral AI
44.84%
111
Gemini 2.0 Flash 001Google
43.61%
112
GPT-4oOpenAI
43.44%
113
Labs Devstral Small 2512Mistral AI
43.18%
114
GPT-4.1 NanoOpenAI
42.72%
115
Grok 4.1 Fast Non ReasoningxAI
42.62%
116
Claude 3.5 HaikuAnthropic
41.92%
117
Gemini 1.5 Pro 002Google
41.72%
118
Claude Haiku 4.5 20251001 ThinkingAnthropic
41.17%
119
Grok 2 1212xAI
38.68%
120
Meta Llama Llama 4 Scout 17B 16E InstructTogether AI
38.54%
121
Mistral Large 2411Mistral AI
37.09%
122
Gemini 1.5 Flash 002Google
36.91%
123
Meta Llama Llama 3.3 70B Instruct TurboTogether AI
36.34%
124
Langston Nim Nvidia Llama 3.3 Nemotron Super 49b V1 42e84561Together AI
36.31%
125
Command A 03 2025Cohere
35.07%
126
Mistral Small 2503Mistral AI
31.82%
127
GPT-4o MiniOpenAI
26.42%
128
Jamba Large 1.6AI21 Labs
22.32%
129
Command R PlusCohere
18.24%
130
Mistral Small 2402Mistral AI
15.78%
131
Jamba Mini 1.6AI21 Labs
9.92%

The published Vals LiveCodeBench mirror snapshot places Claude Fable 5 first at 89.78%. The third row is 1.29 points behind. The broader top-10 range is 2.47 points, so many of the published results sit in a relatively narrow band.

131 models have been evaluated on Vals LiveCodeBench mirror. The benchmark falls in the External benchmark mirrors category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. Vals LiveCodeBench mirror is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About Vals LiveCodeBench mirror

Year

2026

Tasks

Coding problem difficulty splits

Format

Accuracy score

Difficulty

Contamination-resistant coding problems

BenchLM keeps this separate from its canonical LiveCodeBench rows because it is a Vals-hosted implementation snapshot.

Vals LiveCodeBenchPublic benchmark source

BenchLM freshness & provenance

Version

Vals LiveCodeBench mirror 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does Vals LiveCodeBench mirror measure?

Vals AI implementation of LiveCodeBench with easy, medium, and hard task splits.

Which model leads the published Vals LiveCodeBench mirror snapshot?

Claude Fable 5 currently leads the published Vals LiveCodeBench mirror snapshot with 89.78% vals livecodebench mirror score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on Vals LiveCodeBench mirror?

131 AI models are included in BenchLM's mirrored Vals LiveCodeBench mirror snapshot, based on the public leaderboard captured on July 23, 2026.

Last updated: July 23, 2026 · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.