Examenos

31.1 · raw 31.13 index

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

53.9% · raw 53.93 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

17% · raw 17.01 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

28.7% · raw 28.73 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

21.8% · raw 21.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

44.5% · raw 44.49 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

88.8% · raw 88.83 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

67.4% · raw 67.38 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

56.1% · raw 56.14 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

90.4% · raw 90.35 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

55.6% · raw 55.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

70% · raw 70 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

50.3% · raw 50.25 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

68.7% · raw 68.71 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-15

Harness: NVIDIA evaluation of the released BF16 post-trained model. Bolt report Table 7: explicit BF16 columns; all other metrics inherit Super parent evaluation setup. GDPVal is a published index, not Elo or percent; its mapping to Table 10 normalization is not asserted. MCP-Atlas uses the official evaluator on 500 public tasks with reasoning on, not the full 1,000-task suite. WebArena uses 812 public tasks, screenshot vision, vLLM temp 1.0/top_p 0.95, 100 steps, internal string/URL/programmatic checks. BrowseComp uses Exa. Super values differ from its main launch run and remain alternatives. Checked 15 October 2026; execution date and selectable model effort unpublished unless explicitly stated.

979 · raw 979 Elo

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-20

Manually observed public GDPval-AA v2.1 leaderboard table, full 298-configuration pool checked 20 October 2026. Exact source configuration: Nemotron 3.5 Bolt 74B-A9B; creator: NVIDIA; source rank 173; source release-month label Oct 2026 is model release metadata, not execution date. Published rounded native Elo 979; literal displayed CI -19 / +19. Native Elo is not a percentage and is not the normalized (Elo−500)/2000 chart quantity. This versioned metric is separate from older generic GDPval-AA Elo and vendor GDPval/GDPval-AA results; no old evidence is overwritten. Execution date, inference precision, unprinted effort budgets and task costs are unpublished in this captured table. Elo is relative to the dated comparison pool; drift from an earlier observation does not prove a new evaluation run. Source does not label an effort; unknown is preserved rather than copying a vendor default. AA exact model-page links to the approved reference checkpoint were verified in prior curated evidence; a BF16 reference link does not prove BF16 inference. Headline: highest published native Elo for this exact approved model; tied values prefer existing featured effort then source order.

21.4 · raw 21.4072054842003 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

-19.2 · raw -19.166666666666668 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

17.1% · raw 17.083333333333332 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

43.7% · raw 43.71859296482412 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

979 · raw 979.03 Elo

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

0% · raw 0 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

39.7% · raw 39.6990740740741 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

69.7% · raw 69.6666666666667 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

24.7% · raw 24.6524559777572 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

4.3% · raw 4.28571428571429 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

2.8% · raw 2.8000000000000003 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

15.5% · raw 15.46088064040565 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2. Partial reward, not strict pass rate.

22.1 · raw 22.1197398443772 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

19.4 · raw 19.3811765264479 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

23.2 · raw 23.1514559754606 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

24.9 · raw 24.8780755395207 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

32 · raw 32.0477554255484 index

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.

911.9 · raw 911.94 Elo

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2. 95% CI 902.14 to 921.73.

944.4 · raw 944.41 Elo

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2. 95% CI 926.72 to 962.11.

825.4 · raw 825.4 Elo

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2. 95% CI 807.24 to 843.55.

24.3% · raw 24.343434343434346 %

Headline · unknown effort · Independent evaluator

Source/record date: 2026-10-17

Direct independent AA exact model page explicitly links https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16. Inference precision, selectable effort and execution dates unpublished; linking the reference BF16 repository does not prove BF16 serving. reasoningTokens metadata is not an inference budget. Index v4.3.2.