Examenos

92.6% · raw 92.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

60.2% · raw 60.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

43.6% · raw 43.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. HLE without tools; the table lists HLE w/ tools separately. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

82.8% · raw 82.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

81.8% · raw 81.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

77.8% · raw 77.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

86.1% · raw 86.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

86.1% · raw 86.1 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max-0902 comparison table via DataCamp; vendor official preferred over Nex-N2.5-Pro peer comparison table (rule a); same value. Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

93% · raw 93 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. BasicAgent, Code-Dev, judged by Opus 4.6, average of 3 runs. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

63.5% · raw 63.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

67.7% · raw 67.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Claude Code, temperature 1.0, top_p 0.95, 256K; problematic tasks corrected. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

67.7% · raw 67.7 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max-0902 comparison table via DataCamp; vendor official preferred over Nex-N2.5-Pro peer comparison table (rule a); same value. Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

67.7% · raw 67.7 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI. Chart labels Qwen-3.8-Max; the value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

86.6% · raw 86.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Claude Code avg@10, 5h, 131,072 output limit. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

86.6% · raw 86.6 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max-0902 comparison table via DataCamp; vendor official preferred over Nex-N2.5-Pro peer comparison table (rule a); same value (Nex row effort unknown, vendor row max). Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

88.8% · raw 88.8 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI. Chart labels Qwen-3.8-Max; the value matches Alibaba’s August table. NaiveAI’s chart labels Qwen-3.8-Max with no snapshot; Alibaba’s August table reads 86.6, so this figure differs. No averaging. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

69% · raw 69 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Average of frontend, webpage and website; Claude Code; gpt-5.4-2026-03-05 judge. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot. Re-pointed 2026-11-01 from datacamp.com, which transcribed this table; value read from the original.

75.1% · raw 75.1 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Harness: Nex-AGI Vision2Web eval (avg of Frontend/Webpage/Website; Gemini-3.5-Flash VLM judge, GLM-5V-Turbo (Claude Code) GUI agent; Nex card footnote 7) — different harness from Alibaba's figure. Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max table via DataCamp (69.0); vendor official preferred (rule a). Alibaba’s August table reads 69.0 with its own harness. No averaging. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

72.5% · raw 72.5 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max-0902 comparison table via DataCamp (73.3); vendor official 0902 figure preferred (rules a+b). 72.5 equals the pre-0902 Qwen3.8-Max (2026-08-03) value in Alibaba's table, so Nex's "Qwen3.8-Max" column appears to be the original snapshot, not 0902. Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

72.5% · raw 72.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

53.4% · raw 53.4 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-09-25: duplicate of headline from Alibaba Qwen3.8-Max-0902 comparison table via DataCamp (64.0); vendor official 0902 figure preferred (rules a+b). 53.4 equals the pre-0902 Qwen3.8-Max (2026-08-03) value in Alibaba's table, so Nex's "Qwen3.8-Max" column appears to be the original snapshot, not 0902. Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

53.4% · raw 53.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

46.7% · raw 46.7 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-11-01: duplicate of headline from Alibaba’s Qwen3.8-Max table; own vendor over a peer’s, rule (a). Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

46.7% · raw 46.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Partial score; binary 19.4.

66.8% · raw 66.8 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table | Demoted 2026-11-01: duplicate of headline from Alibaba’s Qwen3.8-Max table; own vendor over a peer’s, rule (a). Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

66.8% · raw 66.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

92.1% · raw 92.1 %

Headline · unknown effort · Peer vendor

Source/record date: 2026-09-08

As reported in Nex-N2.5-Pro model card comparison table Nex labels its column Qwen3.8-Max with no snapshot; this value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

56.6% · raw 56.6 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI. Chart labels Qwen-3.8-Max; the value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

56.6% · raw 56.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Highest of Claude Code and mini-SWE-agent.

55.9% · raw 55.9 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI. Chart labels Qwen-3.8-Max; the value matches Alibaba’s August table. Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

55.9% · raw 55.9 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Claude Code.

73.5% · raw 73.5 %

Alternative · unknown effort · Peer vendor

Source/record date: 2026-09-27

As reported by NaiveAI. Chart labels Qwen-3.8-Max; the value matches Alibaba’s August table. | Demoted 2026-11-01: duplicate of headline from Alibaba’s Qwen3.8-Max table; own vendor over a peer’s, rule (a). Moved 2026-11-01 from qwen3.8-max-0902: Alibaba’s tables give this value to Qwen3.8-Max, not the 0902 snapshot.

73.5% · raw 73.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Claude Code, MEAN@5, dominance recomputed.

19.4% · raw 19.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Binary score; partial 46.7.

41% · raw 41 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Claude Code, 5h.

74.8% · raw 74.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. In-house benchmark.

67.7% · raw 67.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

70.2% · raw 70.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. v1.1, 87 tasks, average of 3 runs, OpenCode.

27% · raw 27 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Pass; Score also reported 52.4.

27.3% · raw 27.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Pass@1 on the 600-task public subset.

81.9% · raw 81.9 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Qwen-Agent harness, average item-F1 over 4 runs.

56.2% · raw 56.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

66.3% · raw 66.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

82.3% · raw 82.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

97.7% · raw 97.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. With CI; without CI 95.2.

82% · raw 82 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Without CI; with CI 91.3.

91.3% · raw 91.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. With CI; without CI 82.0.

80.4% · raw 80.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

84.5% · raw 84.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

85.3% · raw 85.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

77.2% · raw 77.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Pass@3; average score 74.8.

51.7% · raw 51.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

79.6% · raw 79.6 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

93.5% · raw 93.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. With CI; without CI 88.4.

92.1% · raw 92.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

74.2% · raw 74.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

68.3% · raw 68.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

88% · raw 88 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

77.8% · raw 77.8 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

75% · raw 75 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

53.2% · raw 53.2 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

85.9% · raw 85.9 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

82.4% · raw 82.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table.

90.4% · raw 90.4 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Table label VideoMME (w/ Sub.), with subtitles.

68.3% · raw 68.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Table label VideoMME v2 (w/ Sub.); the table does not say whether audio was used.

87% · raw 87 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. Table label Dense200, under visual grounding.

80.7% · raw 80.7 %

Headline · unknown effort · Own vendor

Source/record date: 2026-08-03

Alibaba’s Qwen3.8-Max table. In-house, Claude Code avg@3, 8h, 32,768 output limit, temperature 1.0, 256K.

11.3% · raw 11.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column. Claude Code avg@5, 10h, max_tokens 128000, temperature 1.0, 1M.

10.5% · raw 10.5 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column. Claude Code, 12h, 64000 output limit, 1M.

39.1% · raw 39.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column. Claude Code, 10h, 1M.

60.3% · raw 60.3 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column.

55.1% · raw 55.1 %

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column. In-house, Claude Code avg@3, 8h, 64000 output limit, temperature 1.0, 1M.

1348 · raw 1348 Elo

Headline · unknown effort · Own vendor

Source/record date: 2026-09-02

Alibaba’s 0902 comparison table, Qwen3.8-Max column. In-house, 145 tasks, Elo.

51.840% · raw 51.84 %

Headline · max effort · Independent evaluator

Source/record date: 2026-09-23

Vals listed this as alibaba/qwen3.8-max on 23 September. Vals now has separate pages for Qwen 3.8 Max (3 August) and Qwen 3.8 Max 0902 (2 September, 59.05 on Vals Index v2.2), so the key points to the base model; Vals publishes no run date. Moved 2026-11-01 from qwen3.8-max-0902.

24% · raw 23.96 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Code Migration, 23.96%, published uncertainty ±4.31 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

28.6% · raw 28.57 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label CyberBench v1.1, 28.57%, published uncertainty ±3.31 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

65.990% · raw 65.99 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label EMB v1.1, 65.99%, published uncertainty ±2.69 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

50.6% · raw 50.59 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Finance Agent (v2), 50.59%, published uncertainty ±0.58 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

47.600% · raw 47.6 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Legal Research Bench, 47.60%, published uncertainty ±3.47 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

85% · raw 84.95 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label MedScribe, 84.95%, published uncertainty ±2.00 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

64% · raw 63.99 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label MortgageTax, 63.99%, published uncertainty ±0.94 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

23.870% · raw 23.87 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label MysteryMechanism, 23.87%, published uncertainty ±2.87 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

58% · raw 58 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label ProofBench v1.1, 58.00%, published uncertainty ±4.96 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

81.8% · raw 81.76 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label SAGE, 81.76%, published uncertainty ±2.08 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

67.1% · raw 67.12 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Public Benefits Bench v1.1, 67.12%, published uncertainty ±1.22 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

32.090% · raw 32.09 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Tax Agent Bench, 32.09%, published uncertainty ±3.37 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

64.700% · raw 64.7 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Vibe Code Bench v1.1, 64.70%, published uncertainty ±5.36 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face. Harness: OpenHands.

68.9% · raw 68.89 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label IOI, 68.89%, published uncertainty ±5.20 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

0% · raw 0 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label ProgramBench, 0.00%, published uncertainty ±0.00 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

85.6% · raw 85.6 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label SWE-bench, 85.60%, published uncertainty ±1.57 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face.

34.340% · raw 34.34 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Terminal-Bench 4.0, 34.34%, published uncertainty ±3.94 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face. Harness: Mini-SWE-agent.

12.860% · raw 12.86 %

Headline · max effort · Independent evaluator

Source/record date: 2026-11-01

Vals exact model page https://www.vals.ai/models/alibaba_qwen3.8-max; source label Terminal-Bench-Science, 12.86%, published uncertainty ±4.03 percentage points. Model page defaults are not inferred vendor product limits. Run execution date unpublished. Vals lists this as Qwen 3.8 Max, released 3 August 2026, with a separate page for Qwen 3.8 Max 0902. Vals calls its weights private; Alibaba published official weights on Hugging Face. Harness: Mini-SWE-agent.

19 fixes · raw 19 fixes

Headline · unknown effort · Independent evaluator

Source/record date: 2026-11-01

Harness: Claude Code / OpenRouter; effort default; one run dated 3 August 2026, before the 0902 snapshot. Bug Hunt marks this row superseded by its Alibaba API runs of 11 and 15 September, which stay on qwen3.8-max-0902. Only verified planted fixes count.