Examenos

35% · raw 35 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-03

GDP.PDF all-pass rate

34.6% · raw 34.6 %

Headline · max effort · Own vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. K3.1 Pro defaults: max, temperature 1.0, top-p 0.95 unless otherwise stated.

33.2% · raw 33.2 %

Headline · max effort · Own vendor

Source/record date: 2026-10-12

Harness: Meta launch evaluation; Vendor cites Surge professional PDF work score. Direct AA GDP.PDF all-pass for Full is 32.6, not the vendor-cited 33.2; keep separate. Source checked 12 October 2026; evaluation date unpublished. Transcribed from model-card page 3 image; max effort except Full SWE-Marathon xhigh.

32% · raw 32 %

Headline · high effort · Own vendor

Source/record date: 2026-09-29

Higher reasoning setting; +3.2 over Opus 5.5 with fallbacks; -0.2 vs Astra 32.2

31% · raw 31 %

Headline · max effort · Peer vendor

Source/record date: 2026-09-20

Harness: StepFun vendor-reported evaluation; All-pass professional document reasoning. Full agent/grader/task-subset settings unpublished unless stated. Source inspected 8 October 2026; evaluation date unknown unless stated. Approved peer, max effort as labelled; not an independent run.

34.2% · raw 34.2 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-12

Harness: As reported by Meta for this peer model; Vendor cites Surge professional PDF work score. Direct AA GDP.PDF all-pass for Full is 32.6, not the vendor-cited 33.2; keep separate. Source checked 12 October 2026; evaluation date unpublished. Transcribed from model-card page 3 image; max effort except Full SWE-Marathon xhigh.

34.2% · raw 34.2 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Table labels max with fallback; chart notes mention possible Claude fallbacks/GPT cyberguards. This is not a guaranteed single-model-only execution. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. As reported by Moonshot; preserve the peer’s own-vendor headline and source-specific setup.

30.8% · raw 30.8 %

Headline · max effort · Own vendor

Source/record date: 2026-10-12

Harness: Meta launch evaluation; Vendor cites Surge professional PDF work score. Direct AA GDP.PDF all-pass for Full is 32.6, not the vendor-cited 33.2; keep separate. Source checked 12 October 2026; evaluation date unpublished. Transcribed from model-card page 3 image; max effort except Full SWE-Marathon xhigh.

30.6% · raw 30.6 %

Headline · max effort · Peer vendor

Source/record date: 2026-10-12

Harness: As reported by Meta for this peer model; Vendor cites Surge professional PDF work score. Direct AA GDP.PDF all-pass for Full is 32.6, not the vendor-cited 33.2; keep separate. Source checked 12 October 2026; evaluation date unpublished. Transcribed from model-card page 3 image; max effort except Full SWE-Marathon xhigh.

30.6% · raw 30.6 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Table labels max with fallback; chart notes mention possible Claude fallbacks/GPT cyberguards. This is not a guaranteed single-model-only execution. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. As reported by Moonshot; preserve the peer’s own-vendor headline and source-specific setup.

26.2% · raw 26.2 %

Headline · max effort · Peer vendor

Source/record date: 2026-09-20

Harness: StepFun vendor-reported evaluation; All-pass professional document reasoning. Full agent/grader/task-subset settings unpublished unless stated. Source inspected 8 October 2026; evaluation date unknown unless stated. Approved peer, max effort as labelled; not an independent run.

27.6% · raw 27.6 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-12

Harness: As reported by Meta for this peer model; Vendor cites Surge professional PDF work score. Direct AA GDP.PDF all-pass for Full is 32.6, not the vendor-cited 33.2; keep separate. Source checked 12 October 2026; evaluation date unpublished. Transcribed from model-card page 3 image; max effort except Full SWE-Marathon xhigh.

27.6% · raw 27.6 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Table labels max with fallback; chart notes mention possible Claude fallbacks/GPT cyberguards. This is not a guaranteed single-model-only execution. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. As reported by Moonshot; preserve the peer’s own-vendor headline and source-specific setup.

24.7% · raw 24.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-07-09

22% · raw 22 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-10-06

Mistral launch figure 14; as reported by Mistral. AA all-pass document benchmark reported in Mistral launch.

22% · raw 22 %

Alternative · max effort · Peer vendor

Source/record date: 2026-09-20

Harness: StepFun vendor-reported evaluation; All-pass professional document reasoning. Full agent/grader/task-subset settings unpublished unless stated. Source inspected 8 October 2026; evaluation date unknown unless stated. Approved peer, max effort as labelled; not an independent run. Existing headline retained; alternate vendor configuration/evidence, no averaging.

22% · raw 22 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-10-19

Harness: Vendor explicitly attributes this figure to Artificial Analysis. It remains a vendor-cited observation; the exact new model was not found in the checked AA catalog. As reported by MiniMax for the exact named peer; not an independent measurement. Existing headline/configuration remains preferred when present. Main figure contains no per-model effort labels. Execution date, exact checkpoint SHA, generation budget and sampling settings are unpublished. Source checked 2026-10-19.

20.8% · raw 20.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-19

Harness: Vendor explicitly attributes this figure to Artificial Analysis. It remains a vendor-cited observation; the exact new model was not found in the checked AA catalog. Own vendor launch figure. Main figure contains no per-model effort labels. Execution date, exact checkpoint SHA, generation budget and sampling settings are unpublished. Source checked 2026-10-19.

18.6% · raw 18.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-10-06

Mistral launch figure 14; as reported by Mistral. AA all-pass document benchmark reported in Mistral launch.

17.2% · raw 17.2 %

Headline · max effort · Peer vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. As reported by Moonshot; preserve the peer’s own-vendor headline and source-specific setup.

15.4% · raw 15.4 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-10-06

Mistral launch figure 14; as reported by Mistral. AA all-pass document benchmark reported in Mistral launch.

14.8% · raw 14.8 %

Headline · high effort · Own vendor

Source/record date: 2026-09-20

Harness: StepFun vendor-reported evaluation; All-pass professional document reasoning. Full agent/grader/task-subset settings unpublished unless stated. Source inspected 8 October 2026; evaluation date unknown unless stated. Step 5 Preview high effort.

12.8% · raw 12.8 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-10-06

Mistral launch figure 14; as reported by Mistral. AA all-pass document benchmark reported in Mistral launch.

12.8% · raw 12.8 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-10-19

Harness: Vendor explicitly attributes this figure to Artificial Analysis. It remains a vendor-cited observation; the exact new model was not found in the checked AA catalog. As reported by MiniMax for the exact named peer; not an independent measurement. Existing headline/configuration remains preferred when present. Main figure contains no per-model effort labels. Execution date, exact checkpoint SHA, generation budget and sampling settings are unpublished. Source checked 2026-10-19.

11.2% · raw 11.2 %

Headline · max effort · Peer vendor

Source/record date: 2026-09-20

Harness: StepFun vendor-reported evaluation; All-pass professional document reasoning. Full agent/grader/task-subset settings unpublished unless stated. Source inspected 8 October 2026; evaluation date unknown unless stated. Approved peer, max effort as labelled; not an independent run.

11.2% · raw 11.2 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-10-19

Harness: Vendor explicitly attributes this figure to Artificial Analysis. It remains a vendor-cited observation; the exact new model was not found in the checked AA catalog. As reported by MiniMax for the exact named peer; not an independent measurement. Existing headline/configuration remains preferred when present. Main figure contains no per-model effort labels. Execution date, exact checkpoint SHA, generation budget and sampling settings are unpublished. Source checked 2026-10-19.

11.2% · raw 11.2 %

Alternative · max effort · Peer vendor

Source/record date: 2026-10-23

Harness: No Python. PDF page images resized to longest edge 1,152 pixels, interleaved with Docling text; Gemini 3.5 Flash medium judge. Strict all-rubric pass rate among fully scored responses; five runs. Published 23 October 2026; checked 23 October 2026. Actual run dates, serving checkpoint hashes and complete compute budgets are unpublished. Vendor and independently verified results remain separate. As reported by Moonshot; preserve the peer’s own-vendor headline and source-specific setup.