Skip to main content
BenchLM

Vals IOI v1 archived leaderboard (IOI v1 (Vals, archived))

We show this table for reference; we do not rank on it.

Data verified 36 confirmed releases in the last 30 daysFollow model changes

Archived Vals AI results on 2024 and 2025 International Olympiad in Informatics programming tasks.

IOI v1 score on IOI v1 (Vals, archived) — August 9, 2026

We mirror the published ioi v1 score view for IOI v1 (Vals, archived). Claude Opus 5 leads the public snapshot at 91.67%, followed by GPT-5.6 Sol (86.67%) and Qwen3.8 Max (73.00%). We do not use these results to rank models overall.

1Proprietary

Claude Opus 5

Anthropic

91.67%
Overall 79.24
2Proprietary

GPT-5.6 Sol

OpenAI

max reasoning

86.67%
Overall 78.19Context 1.05M
3Open weights

Qwen3.8 Max

Alibaba

73.00%
Overall 71.44Context 1M

62 modelsCodingCurrentDisplay onlyUpdated August 9, 2026

IOI v1 score table (62 models)

Score
1
Claude Opus 5Anthropic · Closed
91.67%
2
GPT-5.6 SolOpenAI · Closedmax reasoning
86.67%
3
Qwen3.8 MaxAlibaba · Open weight
73.00%
4
GPT-5.6 LunaOpenAI · Closedmax reasoning
72.92%
5
Claude Fable 5Anthropic · Closed
72.25%
6
GPT-5.4OpenAI · Closedxhigh reasoning
67.83%
7
GPT-5.6 TerraOpenAI · Closedxhigh reasoning
65.25%
8
GPT-5.2OpenAI · Closedxhigh reasoning
54.83%
9
Muse Spark 1.2Meta · Closed
49.50%
10
Claude Opus 4.7Anthropic · Closed
47.08%
11
Qwen3.7 MaxAlibaba · Closed
46.75%
12
GPT-5.3 CodexOpenAI · Closedxhigh reasoning
43.83%
13
Gemini 3 Flash PreviewGooglehigh reasoning
39.08%
14
Gemini 3 Pro PreviewGooglehigh reasoning
38.83%
15
DeepSeek V4 Pro 0813DeepSeek · Open weightmax reasoning
35.83%
16
Grok 4.20 0309 ReasoningxAI
30.17%
17
Gemini 3.5 Flash-LiteGoogle · Closedhigh reasoning
26.17%
18
Grok 4 0709xAI
26.17%
19
Claude Opus 4.5Anthropic · Closed
23.58%
20
GLM 5 ThinkingZhipu AI
22.00%
21
GPT-5.1OpenAI · Closedhigh reasoning
21.50%
22
GPT-5.1-Codex-MaxOpenAI · Closedhigh reasoning
21.42%
23
Claude Opus 4.5 ThinkingAnthropic · Closed
20.25%
24
GPT-5OpenAIhigh reasoning
20.00%
25
Claude Sonnet 4.5 ThinkingAnthropic · Closed
18.33%
26
Kimi K2.5 ThinkingMoonshot AI
17.67%
27
Gemini 2.5 ProGoogle · Closed
17.08%
28
Qwen3 MaxAlibaba · Closed
15.67%
29
Grok 4.3xAI · Closedhigh reasoning
15.33%
30
GPT-5.4 nanoOpenAI · Closedhigh reasoning
15.25%
31
DeepSeek V3p2Fireworks AInone reasoning
14.42%
32
Qwen3 MaxAlibaba · Closed
13.75%
33
Claude Opus 4.1Anthropic
12.52%
34
Grok 4 Fast (Reasoning)xAI · Closed
11.50%
35
DeepSeek V3p2 ThinkingFireworks AIhigh reasoning
10.67%
36
GPT-5 CodexOpenAIhigh reasoning
9.75%
37
Qwen3 Max PreviewAlibaba
7.75%
38
Grok 4.1 Fast Non ReasoningxAI
7.67%
39
GLM-4.7Z.AI · Open weight
7.58%
40
GPT-5 miniOpenAI · Closedhigh reasoning
6.75%
41
MiniMax M2.5MiniMax · Closed
6.67%
42
Claude Sonnet 4Anthropic
6.50%
43
GPT-5.4 miniOpenAI · Closedxhigh reasoning
6.42%
44
Claude Haiku 4.5 ThinkingAnthropic · Closed
6.17%
45
MiniMax M2.7MiniMax · Open weight
4.92%
46
O4 MiniOpenAIhigh reasoning
4.83%
47
Claude Sonnet 4 20250514 ThinkingAnthropic
4.58%
48
GLM-4.6Z.AI · Open weight
4.33%
49
Grok Code Fast 1xAI · Closed
4.33%
50
Mistral Large 2512Mistral AI
4.00%
51
Grok 4 Fast Non ReasoningxAI
3.83%
52
GPT-5.1-CodexOpenAI · Closedhigh reasoning
3.67%
53
Grok 4.1 Fast (Reasoning)xAI · Closed
3.08%
54
GLM-4.5Z.AI · Closed
2.92%
55
Gemini 2.5 FlashGoogle · Closed
2.61%
56
Labs Devstral Small 2512Mistral AI
2.50%
57
MiniMax M2.1MiniMax
2.33%
58
DeepSeek V3 0324Fireworks AI
1.67%
59
Moonshotai Kimi K2 InstructTogether AI
1.25%
60
Magistral Medium 2509Mistral AI
0.67%
61
Devstral 2512Mistral AI
0.67%
62
Qwen3 235b A22bFireworks AI
0.00%

How IOI v1 (archived) is shown here

BenchLM mirrors the public Vals AI IOI v1 (archived) leaderboard captured from https://www.vals.ai/benchmarks/ioi-v1 and updated by Vals on August 9, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

IOI v1 (archived) 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.

Snapshot

62 Vals rows3 task viewspublic datasetTasks: Overall, IOI 2024, IOI 2025Display only
IOI v1 (archived) on Vals AIVals methodologyVals home

The published IOI v1 (Vals, archived) snapshot places Claude Opus 5 first at 91.67%. The third row is 18.67 points behind. The broader top-10 range is 44.58 points, so the table still separates the published systems.

62 models have been evaluated on IOI v1 (Vals, archived). The benchmark falls in the Coding category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. IOI v1 (Vals, archived) is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About IOI v1 (Vals, archived)

Year

2026

Tasks

IOI 2024 and 2025 programming tasks

Format

Accuracy score

Difficulty

Olympiad programming

Vals stopped running IOI v1 on new model releases and preserves its prior results. This display-only version-1 table stays separate from the current IOI v2 task set, which also includes IOI 2026.

Vals IOI v1Public benchmark source

Freshness and provenance

Version

IOI v1 (Vals, archived) 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.

Questions

What does IOI v1 (Vals, archived) measure?

Archived Vals AI results on 2024 and 2025 International Olympiad in Informatics programming tasks.

Which model leads the published IOI v1 (Vals, archived) snapshot?

Claude Opus 5 currently leads the published IOI v1 (Vals, archived) snapshot with 91.67% ioi v1 score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on IOI v1 (Vals, archived)?

The August 9, 2026 snapshot contains 62 AI models.

Last updated: August 9, 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.