Benchmark Almanac

Codeforces rating

DeepSeek V4.1 Flash 3,471 leads since Sep 2026 In use

Model solves recent Codeforces contest problems. An Elo-style rating is computed from simulated contest performance.

No standard public harness. Each lab's rating method differs, so numbers are not comparable across cards.

Made by Codeforces (problems); each lab runs its own harness Released 2024 Competitive programming Last in a launch post Sep 2026 Last model added Sep 2026
Level
Competition The strongest students, in olympiad-style contests
Scored
Checked automatically A program checks the result against an answer key, unit tests or the expected final state.
Answers
Code A program or patch run against tests.
41
Models scored in lab cards
28
Launch posts list it, from 7 labs
33
System cards report it, from 9 labs

DeepSeek V4.1 Flash leads with 3,471 among models released by Sep 2026

Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.

The newest model with a score here came out in Sep 2026, and nobody has reported a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

41 models. Scroll sideways for the rest.

  1. 3,471 DeepSeek V4.1 Flash
  2. 3,455 Gemini 3 Deep Think
  3. 3,206 DeepSeek-V4-Pro
  4. 3,052 DeepSeek-V4-Flash
  5. 3,020 Seed2.0 Pro
  6. 2,748 o3-pro
  7. 2,719 o4-mini
  8. 2,706 o3
  9. 2,701 DeepSeek-V3.2-Speciale
  10. 2,622 gpt-oss-120b
  11. 2,516 gpt-oss-20b
  12. 2,386 DeepSeek-V3.2
  13. 2,320 Nemotron-Cascade 2 30B-A3B
  14. 2,150 Gemma 4 31B IT
  15. 2,134 Qwen3-235B-A22B-Thinking
  16. 2,130 o3-mini
  17. 2,121 DeepSeek-V3.2-Exp
  18. 2,100 Qwen3.5-122B-A10B
  19. 2,100 qwen3-coder-next
  20. 2,091 DeepSeek-V3.1
  21. 2,088 Ring-1T
  22. 2,071 Qwen3-Next-80B-A3B
  23. 2,056 Qwen3-235B-A22B
  24. 2,046 DeepSeek-V3.1-Terminus
  25. 2,029 DeepSeek-R1
  26. 1,977 Qwen3-32B
  27. 1,974 Qwen3-30B-A3B
  28. 1,964 Qwen3-VL-235B-A22B
  29. 1,785 Qwen3-8B
  30. 1,766 Qwen3-14B
  31. 1,736 Phi-4-reasoning
  32. 1,723 Phi-4-Reasoning-plus
  33. 1,718 Gemma 4 26B A4B
  34. 1,691 DeepSeek-R1-Distill-Qwen-32B
  35. 1,671 Qwen3-4B
  36. 1,633 DeepSeek-R1-Distill-Llama-70B
  37. 1,481 DeepSeek-R1-Distill-Qwen-14B
  38. 1,429 DiffusionGemma 26B A4B
  39. 1,258 o1-preview
  40. 954 DeepSeek-R1-Distill-Qwen-1.5B
  41. 392 GPT-4
  1. 1 DeepSeek V4.1 Flash 3,471
  2. 2 Gemini 3 Deep Think 3,455
  3. 3 DeepSeek-V4-Pro 3,206
  4. 4 DeepSeek-V4-Flash 3,052
  5. 5 Seed2.0 Pro 3,020
  6. 6 o3-pro 2,748
  7. 7 o4-mini 2,719
  8. 8 o3 2,706
  9. 9 DeepSeek-V3.2-Speciale 2,701
  10. 10 gpt-oss-120b 2,622
  11. 11 gpt-oss-20b 2,516
  12. 12 DeepSeek-V3.2 2,386
  13. 13 Nemotron-Cascade 2 30B-A3B 2,320
  14. 14 Gemma 4 31B IT 2,150
  15. 15 Qwen3-235B-A22B-Thinking Jul 2025 2,134
  16. 16 o3-mini 2,130
  17. 17 DeepSeek-V3.2-Exp 2,121
  18. 18 Qwen3.5-122B-A10B 2,100
  19. 19 qwen3-coder-next 2,100
  20. 20 DeepSeek-V3.1 2,091
  21. 21 Ring-1T 2,088
  22. 22 Qwen3-Next-80B-A3B 2,071
  23. 23 Qwen3-235B-A22B 2,056
  24. 24 DeepSeek-V3.1-Terminus 2,046
  25. 25 DeepSeek-R1 2,029
  26. 26 Qwen3-32B 1,977
  27. 27 Qwen3-30B-A3B 1,974
  28. 28 Qwen3-VL-235B-A22B 1,964
  29. 29 Qwen3-8B 1,785
  30. 30 Qwen3-14B 1,766
  31. 31 Phi-4-reasoning 1,736
  32. 32 Phi-4-Reasoning-plus 1,723
  33. 33 Gemma 4 26B A4B 1,718
  34. 34 DeepSeek-R1-Distill-Qwen-32B 1,691
  35. 35 Qwen3-4B 1,671
  36. 36 DeepSeek-R1-Distill-Llama-70B 1,633
  37. 37 DeepSeek-R1-Distill-Qwen-14B 1,481
  38. 38 DiffusionGemma 26B A4B 1,429
  39. 39 o1-preview 1,258
  40. 40 DeepSeek-R1-Distill-Qwen-1.5B 954
  41. 41 GPT-4 Mar 2023 392
#ModelScoreSettingDateSource
1 DeepSeek V4.1 Flash 3,471 Sep 2026 Model card
2 Gemini 3 Deep Think 3,455 Feb 2026 Model card
3 DeepSeek-V4-Pro 3,206 Apr 2026 Model card
4 DeepSeek-V4-Flash 3,052 Apr 2026 Model card
5 Seed2.0 Pro 3,020 Feb 2026 Model card
6 o3-pro 2,748 Jun 2025 Model card
7 o4-mini 2,719 Apr 2025 Model card
8 o3 2,706 Apr 2025 Model card
9 DeepSeek-V3.2-Speciale 2,701 Dec 2025 Model card
10 gpt-oss-120b 2,622 Aug 2025 Model card
11 gpt-oss-20b 2,516 Aug 2025 Model card
12 DeepSeek-V3.2 2,386 Dec 2025 Model card
13 Nemotron-Cascade 2 30B-A3B 2,320 Mar 2026 Model card
14 Gemma 4 31B IT 2,150 Apr 2026 Model card
15 Qwen3-235B-A22B-Thinking Jul 2025 2,134 Jul 2025 Model card
16 o3-mini 2,130 Jan 2025 Model card
17 DeepSeek-V3.2-Exp 2,121 Sep 2025 Model card
18 Qwen3.5-122B-A10B 2,100 Feb 2026 Model card
19 qwen3-coder-next 2,100 Feb 2026 Model card
20 DeepSeek-V3.1 2,091 Aug 2025 Model card
21 Ring-1T 2,088 Oct 2025 Model card
22 Qwen3-Next-80B-A3B 2,071 Sep 2025 Model card
23 Qwen3-235B-A22B 2,056 Apr 2025 Model card
24 DeepSeek-V3.1-Terminus 2,046 Sep 2025 Model card
25 DeepSeek-R1 2,029 Jan 2025 Model card
26 Qwen3-32B 1,977 May 2025 Model card
27 Qwen3-30B-A3B 1,974 May 2025 Model card
28 Qwen3-VL-235B-A22B 1,964 Nov 2025 Model card
29 Qwen3-8B 1,785 May 2025 Model card
30 Qwen3-14B 1,766 May 2025 Model card
31 Phi-4-reasoning 1,736 Apr 2025 Model card
32 Phi-4-Reasoning-plus 1,723 Apr 2025 Model card
33 Gemma 4 26B A4B 1,718 Apr 2026 Model card
34 DeepSeek-R1-Distill-Qwen-32B 1,691 Jan 2025 Model card
35 Qwen3-4B 1,671 May 2025 Model card
36 DeepSeek-R1-Distill-Llama-70B 1,633 Jan 2025 Model card
37 DeepSeek-R1-Distill-Qwen-14B 1,481 Jan 2025 Model card
38 DiffusionGemma 26B A4B 1,429 Jun 2026 Model card
39 o1-preview 1,258 Sep 2024 Model card
40 DeepSeek-R1-Distill-Qwen-1.5B 954 Jan 2025 Model card
41 GPT-4 Mar 2023 392 Mar 2023 Model card

8 models have held number one since its release in Jan 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

OpenAI DeepSeek Google DeepMind
Lab claims Not tested since Sep 2026
DeepSeek V4.1 Flash 3,471

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Reported by the lab in its model card
0 1,000 2,000 3,000 4,000 2023 2024 2025 2026 0 1,000 2,000 3,000 4,000 2023 2024 2025 2026

2 labs put it in a launch post in the last six months

Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.

OpenAI
DeepSeek
Moonshot AI
ByteDance Seed
Qwen
Ant Group
Google DeepMind
DateLaunch postScoreAs the post writes it
10 Sep 2026 DeepSeek V4.1 Flash 3471 Codeforces (Rating)
10 Jun 2026 DiffusionGemma 26B A4B 1429 Codeforces ELO
24 Apr 2026 DeepSeek-V4-Pro 3206 Codeforces (Rating)
2 Apr 2026 Gemma 4 26B A4B 1718 Codeforces ELO
2 Apr 2026 Gemma 4 31B IT 2150 Codeforces ELO
24 Feb 2026 Qwen3.5-122B-A10B 2100 CodeForces (Qwen query set)
14 Feb 2026 Seed2.0 Pro 3020 Codeforces (no tool)
12 Feb 2026 Gemini 3 Deep Think 3455 Codeforces Elo
1 Dec 2025 DeepSeek-V3.2 2386 Codeforces (Rating)
10 Oct 2025 Ring-1T 2088 CodeForces (rating)
29 Sep 2025 DeepSeek-V3.2-Exp 2121 Codeforces
22 Sep 2025 DeepSeek-V3.1-Terminus 2046 Codeforces
11 Sep 2025 Qwen3-Next-80B-A3B 2071 CFEval
21 Aug 2025 DeepSeek-V3.1 2091 Codeforces-Div1 (Rating)
5 Aug 2025 gpt-oss-20b 2516 Codeforces (with tools)
5 Aug 2025 gpt-oss-120b 2622 Codeforces Competition code, gpt-oss-120b (with tools)
25 Jul 2025 Qwen3-235B-A22B-Thinking 2134 CFEval
10 Jun 2025 o3-pro 2748 Competition Code (Codeforces), pass@1 Elo, updated question set
28 May 2025 DeepSeek-R1 1930 Codeforces-Div1 (Rating)
29 Apr 2025 Qwen3-235B-A22B 2056 CodeForces (Rating / Percentile)
16 Apr 2025 o4-mini 2719 Codeforces (with terminal)
16 Apr 2025 o3 2706 Codeforces Competition Code, o3 (with terminal)
10 Apr 2025 Seed-Thinking-v1.5 55 Codeforces pass@8 55.0% (avg@8 36.3%)
31 Jan 2025 o3-mini 2130 Competition Code (Codeforces) ELO
20 Jan 2025 Kimi k1.5 94 Codeforces (Percentile), long-CoT
20 Jan 2025 DeepSeek-R1 2029 Codeforces (Rating)
26 Dec 2024 DeepSeek-V3 51.6 Codeforces (Percentile)
12 Sep 2024 o1-preview 1258 CodeForces Elo

33 system cards report it, from 9 labs

Besides launch posts, we read the system cards, model cards and technical reports labs publish with a release. They report far more, including safety and frontier-risk tests.

DateDocumentScoreAs the document writes itSection
10 Sep 2026 DeepSeek V4.1 Flash technical report 3471 Codeforces (Rating), internal benchmark (Table 3) Capabilities
31 Jul 2026 DiffusionGemma 26B A4B technical report 1429 Codeforces ELO, TD, 959 no-think, 1569 AR (Table 3) Capabilities
24 Jul 2026 Gemma 4 31B IT technical report 2150 Codeforces Elo, thinking (Table 5) Capabilities
24 Jul 2026 Gemma 4 26B A4B technical report 1718 Codeforces Elo, thinking (Table 5) Capabilities
26 Apr 2026 DeepSeek-V4-Pro technical report 3206 Codeforces (Rating), internal benchmark of 14 Div. 1 contests (May to Nov 2025), DS-V4-Pro Max (Table 6) Capabilities
26 Apr 2026 DeepSeek-V4-Flash technical report 3052 Codeforces (Rating), internal benchmark, DeepSeek-V4-Flash Max (Table 7) Capabilities
19 Mar 2026 Nemotron-Cascade 2 30B-A3B technical report 2320 Codeforces ELO, simulated on 40 Div.1/Div.2 rounds 2501 to 2507, percentile 99.6 (Table 6). TIR 2345 Capabilities
28 Feb 2026 qwen3-coder-next technical report 2100 Codeforces rating (Table 6) Capabilities
14 Feb 2026 Seed2.0 Pro model card 3020 Codeforces (no tool), June to December 2025 problems (Table 3) Capabilities
1 Dec 2025 DeepSeek-V3.2 technical report 2386 Codeforces (Rating), DeepSeek-V3.2 Thinking (Table 2) Capabilities
1 Dec 2025 DeepSeek-V3.2-Speciale technical report 2701 CodeForces (Rating), DeepSeek-V3.2-Speciale (Table 3) Capabilities
26 Nov 2025 Qwen3-VL-235B-A22B technical report 1964 CFEval (Table 6) Capabilities
21 Oct 2025 Ring-1T technical report 2088 CodeForces (rating), 14 Div. 2 contests, max attainable 2209 (Table 1). Percentile 97.85 Capabilities
29 Sep 2025 DeepSeek-V3.2-Exp technical report 2121 Codeforces-Div1 (Rating), DeepSeek-V3.2-Exp (Table 1) Capabilities
5 Aug 2025 gpt-oss-20b model card 2230 Codeforces (no tools), Elo, high reasoning (with tools 2516) (Table 3) Capabilities
5 Aug 2025 gpt-oss-120b model card 2463 Codeforces (no tools), Elo, high reasoning (with tools 2622) (Table 3) Capabilities
14 May 2025 Qwen3-8B technical report 1785 CodeForces (Rating / Percentile), thinking mode, cell 1785 / 95.6% (Table 17). Non-thinking 1110 / 52.4% (Table 18) Capabilities
14 May 2025 Qwen3-4B technical report 1671 CodeForces (Rating / Percentile), thinking mode, cell 1671 / 92.8% (Table 17). Non-thinking 842 / 33.7% (Table 18) Capabilities
14 May 2025 Qwen3-32B technical report 1977 CodeForces (Rating / Percentile), thinking mode, cell 1977 / 97.7% (Table 13). Non-thinking 1353 / 71.0% (Table 14) Capabilities
14 May 2025 Qwen3-30B-A3B technical report 1974 CodeForces (Rating / Percentile), thinking mode, cell 1974 / 97.7% (Table 15). Non-thinking 1267 / 64.1% (Table 16) Capabilities
14 May 2025 Qwen3-235B-A22B technical report 2056 CodeForces (Rating / Percentile), thinking mode, cell 2056 / 98.2% (Table 11). Non-thinking 1387 / 75.7% (Table 12) Capabilities
14 May 2025 Qwen3-14B technical report 1766 CodeForces (Rating / Percentile), thinking mode, cell 1766 / 95.3% (Table 15). Non-thinking 1200 / 58.6% (Table 16) Capabilities
30 Apr 2025 Phi-4-reasoning technical report 1736 Codeforces rating (Table 1) Capabilities
30 Apr 2025 Phi-4-Reasoning-plus technical report 1723 Codeforces rating (Table 1) Capabilities
10 Apr 2025 Seed-Thinking-v1.5 technical report 55 Codeforces pass@8 55.0%, avg@8 36.3%, a solve rate on recent contests, not a rating (Table 2) Capabilities
20 Jan 2025 Kimi k1.5 technical report Codeforces (Percentile) 94, long-CoT, percentile only and no rating (Table 2) Capabilities
20 Jan 2025 DeepSeek-R1 technical report 2029 Codeforces (Rating), percentile 96.3, DeepSeek-R1 (Table 8, same as final column of Table 3) Capabilities
20 Jan 2025 DeepSeek-R1-Distill-Qwen-32B technical report 1691 CodeForces rating, DeepSeek-R1-Distill-Qwen-32B (Table 15) Capabilities
20 Jan 2025 DeepSeek-R1-Distill-Qwen-14B technical report 1481 CodeForces rating, DeepSeek-R1-Distill-Qwen-14B (Table 15) Capabilities
20 Jan 2025 DeepSeek-R1-Distill-Qwen-1.5B technical report 954 CodeForces rating, DeepSeek-R1-Distill-Qwen-1.5B (Table 15) Capabilities
20 Jan 2025 DeepSeek-R1-Distill-Llama-70B technical report 1633 CodeForces rating, DeepSeek-R1-Distill-Llama-70B (Table 15) Capabilities
26 Dec 2024 DeepSeek-V3 technical report Codeforces (Percentile) 51.6, percentile only and no rating, DeepSeek-V3 (Table 6) Capabilities
14 Mar 2023 GPT-4 technical report 392 Codeforces Rating, 392 (below 5th percentile), average equilibrium rating over 10 contests (Table 1, GPT-4 with vision) Capabilities

More in Coding

All 243 in Coding
  • SWE-bench Verified In use 500 human-validated GitHub issues from 12 Python repos. Agent writes a patch that must pass hidden tests. Claude Opus 4.7 83.5%
  • SWE-bench Multilingual In use 300 GitHub-issue tasks across nine programming languages in SWE-bench format. Patch graded by repository tests. Claude Opus 5.5 93.9%
  • LiveCodeBench v6 In use Competitive programming problems collected after model cutoffs from LeetCode, AtCoder, Codeforces. Scored pass@1 on hidden tests. DeepSeek-V4-Pro 93.5%
  • DeepSWE In use 113 original long-horizon engineering tasks across 91 repos and five languages, written from scratch. Hand-written verifiers accept alternative solutions. GPT-6 Astra 74.1%
  • SciCode In use Scientist-curated research coding problems split into subproblems across natural sciences. Code checked against test cases. Seed2.1 Pro 59.8%
  • NL2Repo-Bench In use From one natural-language requirements document and an empty workspace, agent builds an installable Python library. Scored by original project's test pass rate. DeepSeek V4.1 Flash 1017 68.2%