GPT-5.4

Data updated 12 Sept 2026

32 published benchmark measures · 8 benchmark families contribute across 5 task areas. 4 capability estimates available in the full profile. See all results ↓ · Compare published benchmarks →

Model evidence summary

Choose a configuration

A model-wide score would mix effort settings. This page summarizes published evidence; only identified configurations receive leaderboard ranks.

Compare measured configurations →

The capability chart summarizes evidence across reported settings, not a runnable configuration or a leaderboard rank.

Performance profile

Capabilities

Adjusted comparison score · 0–100

Filled points are supported; hollow points are preliminary. Lines stop at unknown capabilities. Exact values and sources follow below.

AgenticNo comparable evidenceUnknown

0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 11/11 scenarios unsupported. Without one publisher: No supported estimate; 15/15 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage; source sensitivity still requires inspection. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Hard reasoning2 families · 1 with independent evidence · Supported65.5

2 core families; 51 direct opponents across 11 labs.

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: 62.6–71.1; 2/7 scenarios unsupported. Without one publisher: 57.7–80.5; 2/16 unsupported. Smoothing check: 62.9–66.2; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

ARC-AGI: 76.6 observed win share
arcprize.org · Source 1
GPQA: 100.0 observed win share
Microsoft · Source 1

Coding3 families · 2 with independent evidence · Supported8.0

3 core families; 31 direct opponents across 10 labs.

Disputed order: matched results across at least two families give the opposite order against 1 peers. Different test mixes and indirect comparisons can cause this. Matched comparison 1

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: 6.8–25.3; 2/7 scenarios unsupported. Without one publisher: 2.2–30.0; 1/20 unsupported. Smoothing check: 5.5–14.5; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts improved on neutral predictions within evaluated coverage. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

SWE-bench Pro: 58.3 observed win share
Microsoft, scale.com · Source 1 Source 2
DeepSWE: 22.2 observed win share
deepswe.datacurve.ai · Source 1
Terminal-Bench: 0.0 observed win share
Moonshot AI · Source 1

Human pref1 families · 1 with independent evidence · Supported36.1

1 core families; 114 direct opponents across 19 labs.

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 1/1 scenarios unsupported. Without one publisher: No supported estimate; 1/1 unsupported. Smoothing check: 28.3–42.9; 0/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: cross-family and cross-publisher predictions could not be evaluated from the available comparison network. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Human preference (Arena): 85.1 observed win share
lmarena.ai · Source 1

Knowledge1 families · 1 with independent evidence · PreliminaryUnknown

1 core families; 51 direct opponents across 12 labs. No connected comparison to the complete reference panel

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 4/4 scenarios unsupported. Without one publisher: No supported estimate; 5/5 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: almost no publisher-held-out comparisons remained estimable. Cross-source validity is unresolved. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

MMLU-Pro: 94.1 observed win share
huggingface.co · Source 1

Multimodal1 families · 0 with independent evidence · Preliminary52.0

1 core families; 2 direct opponents across 2 labs. Needs broader benchmark and opponent support

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 8/8 scenarios unsupported. Without one publisher: No supported estimate; 8/8 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: limited publisher-held-out coverage. Cross-source validity is unresolved. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

CharXiv: 100.0 observed win share
Moonshot AI · Source 1

Long contextNo comparable evidenceUnknown

0 core families; 0 direct opponents across 0 labs. No admitted core comparisons

Estimated against the same complete panel: gpt-5.6-sol claude-opus-4-8 gemini-3.1-pro . Indirect comparisons assume performance can be summarized by one strength within this capability.

Refit without one family: No supported estimate; 5/5 scenarios unsupported. Without one publisher: No supported estimate; 5/5 unsupported. Smoothing check: No supported estimate; 3/3 unsupported. These are sensitivity checks, not confidence intervals.

Audit: family and publisher holdouts underperformed neutral predictions. Treat this as an exploratory estimate. Validation audit.

Matched family win shares below describe source evidence; they are not averaged to produce the adjusted score.

Scores estimate outcomes against a shared reference panel; they are not accuracy percentages. Sparse or disconnected evidence cannot qualify an overall profile. Open a capability to inspect its evidence.

Compare 5 effort levels across 51 benchmark/harness combinations →

Reported effort · XHigh + unspecified

Settings reported in this model's published benchmark results, including results outside the aggregate. Effort names are provider-specific. These are not API defaults or equal compute budgets.

XHigh: 28 observations
Not specified: 8 observations

Mixed settings means multiple settings occur in the evidence. Best across efforts means the source selected its best reported result across settings; it does not mean Max. Unspecified settings stay unknown. This model-summary chart combines reported settings. The leaderboard keeps identified configurations separate and excludes unknown effort.

Inspect each result and its source ↓ · Download effort evidence

Compare capability profiles →

Score contributions and missing evidence

8 contributing families across 5 capabilities. Fixed reference panels do not change when the catalog expands.

Capability is fitted jointly across families. Capability estimates below describe different task areas; their weighted sum is not the Capability score.

Results without reviewed compatibility or a reference match remain in the raw evidence below. Coverage counts only contributing results.

Model information & shareable badge

Lab: OpenAI
Catalog status: active
Availability: Public provider catalog; account and region restrictions may apply
Family: gpt-5.4
Released: —
Context: —
License: proprietary
Model card: https://developers.openai.com/api/docs/models/all
Default Capability family coverage: /badge/gpt-5.4.svg

Benchmark scores & sources

Original results, evaluation harnesses, and evidence behind this model.

Benchmark	Bucket	Score	Harness	Evidence	Source-recorded date
AIME 2026	Hard reasoning	99.2% Reported settings & source Kimi K2.6 card, AIME 2026. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Reasoning: 98304 generation tokens; HLE full set. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, AIME 2026, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
APEX-Agents	Agentic	33.3% Reported settings & source Kimi K2.6 card, APEX-Agents. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, APEX-Agents, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
ARC-AGI-2	Hard reasoning	74%	ARC Prize verified Effort: Not specified contributes to capability	official board	2026-03-04
BabyVision	Supporting evidence	49.7% Reported settings & source Kimi K2.6 card, BabyVision. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, BabyVision, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
BabyVision (w/ python)	Supporting evidence	80.2% Reported settings & source Kimi K2.6 card, BabyVision (w/ python). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, BabyVision (w/ python), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
BrowseComp	Agentic	82.7% Reported settings & source Kimi K2.6 card, BrowseComp. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, BrowseComp, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
BrowseComp (Agent Swarm)	Agentic	82.7% Reported settings & source Kimi K2.6 card, BrowseComp (Agent Swarm). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, BrowseComp (Agent Swarm), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
CharXiv (RQ)	Multimodal	82.8% Reported settings & source Kimi K2.6 card, CharXiv (RQ). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, CharXiv (RQ), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh contributes to capability	lab self-report	Reviewed 2026-09-12
CharXiv (RQ) (w/ python)	Supporting evidence	90% Reported settings & source Kimi K2.6 card, CharXiv (RQ) (w/ python). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, CharXiv (RQ) (w/ python), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
Claw Eval (pass@3)	Supporting evidence	78.4% Reported settings & source Kimi K2.6 card, Claw Eval (pass@3). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, Claw Eval (pass@3), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
Claw Eval (pass^3)	Supporting evidence	60.3% Reported settings & source Kimi K2.6 card, Claw Eval (pass^3). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, Claw Eval (pass^3), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
DeepSearchQA (accuracy)	Supporting evidence	63.7% Reported settings & source Kimi K2.6 card, DeepSearchQA (accuracy). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, DeepSearchQA (accuracy), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
DeepSearchQA (f1-score)	Supporting evidence	78.6% Reported settings & source Kimi K2.6 card, DeepSearchQA (f1-score). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, DeepSearchQA (f1-score), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
DeepSWE v1.1	Coding	51.8%	DeepSWE v1.1 reported Effort: Not specified contributes to capability	official board	2026-09-03
GPQA Diamond · source release snapshot; version not specified	Hard reasoning	92.8% Reported settings & source MAI avg 4 runs, temperature 1, top-p .97; simple ReAct bash/string-replace; Terminal-Bench ignores predefined timeouts. Comparator configurations from cited reports. First-party reported result; comparator results retain the source evaluation setup. MAI-Thinking-1 technical report · Table 11, PDF page 53 · reviewed 2026-09-06	Published configuration Effort: Not specified contributes to capability	lab self-report	Reviewed 2026-09-06
GPQA-Diamond	Supporting evidence	92.8% Reported settings & source Kimi K2.6 card, GPQA-Diamond. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Reasoning: 98304 generation tokens; HLE full set. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, GPQA-Diamond, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
HLE-Full	Hard reasoning	39.8% Reported settings & source Kimi K2.6 card, HLE-Full. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Reasoning: 98304 generation tokens; HLE full set. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, HLE-Full, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
HLE-Full (w/ tools)	Hard reasoning	52.1% Reported settings & source Kimi K2.6 card, HLE-Full (w/ tools). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, HLE-Full (w/ tools), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
HMMT 2026 (Feb)	Supporting evidence	97.7% Reported settings & source Kimi K2.6 card, HMMT 2026 (Feb). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Reasoning: 98304 generation tokens; HLE full set. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, HMMT 2026 (Feb), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
IMO-AnswerBench	Supporting evidence	91.4% Reported settings & source Kimi K2.6 card, IMO-AnswerBench. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Reasoning: 98304 generation tokens; HLE full set. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, IMO-AnswerBench, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
LMArena Text Arena	Human pref	1466	LMArena Text Effort: Not specified contributes to capability	official board	2026-09-11
MathVision	Supporting evidence	92% Reported settings & source Kimi K2.6 card, MathVision. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, MathVision, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
MathVision (w/ python)	Supporting evidence	96.1% Reported settings & source Kimi K2.6 card, MathVision (w/ python). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, MathVision (w/ python), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
MCPMark	Supporting evidence	62.5% Reported settings & source Kimi K2.6 card, MCPMark. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, MCPMark, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
MMLU-Pro	Knowledge	87.5%	MMLU-Pro reported Effort: Not specified contributes to capability	official board	2026-09-11
MMMU-Pro	Multimodal	81.2% Reported settings & source Kimi K2.6 card, MMMU-Pro. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, MMMU-Pro, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
MMMU-Pro (w/ python)	Supporting evidence	82.1% Reported settings & source Kimi K2.6 card, MMMU-Pro (w/ python). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, MMMU-Pro (w/ python), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
OSWorld-Verified	Agentic	75% Reported settings & source Kimi K2.6 card, OSWorld-Verified. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, OSWorld-Verified, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
SciCode	Coding	56.6% Reported settings & source Kimi K2.6 card, SciCode. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Coding: average of ten runs; Terminal uses Terminus2 preserve-thinking; SWE uses in-house SWE-agent. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, SciCode, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
SWE-bench Pro	Coding	59.1%	SWE-bench Pro reported Effort: Not specified contributes to capability	official board	2026-04-08
SWE-Bench Pro	Coding	57.7% Reported settings & source Kimi K2.6 card, SWE-Bench Pro. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Coding: average of ten runs; Terminal uses Terminus2 preserve-thinking; SWE uses in-house SWE-agent. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, SWE-Bench Pro, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
SWE-bench Pro · source release snapshot; version not specified	Coding	57.7% Reported settings & source MAI avg 4 runs, temperature 1, top-p .97; simple ReAct bash/string-replace; Terminal-Bench ignores predefined timeouts. Comparator configurations from cited reports. First-party reported result; comparator results retain the source evaluation setup. MAI-Thinking-1 technical report · Table 11, PDF page 53 · reviewed 2026-09-06	Published configuration Effort: Not specified contributes to capability	lab self-report	Reviewed 2026-09-06
Terminal-Bench 2.0 · 2.0	Coding	75.1% Reported settings & source MAI avg 4 runs, temperature 1, top-p .97; simple ReAct bash/string-replace; Terminal-Bench ignores predefined timeouts. Comparator configurations from cited reports. First-party reported result; comparator results retain the source evaluation setup. MAI-Thinking-1 technical report · Table 11, PDF page 53 · reviewed 2026-09-06	Published configuration Effort: Not specified	lab self-report Microsoft report Table 11 and section 4.1: MAI Terminal-Bench removes timeouts and uses a minimal ReAct harness; comparator values are cited from other model releases. A common evaluation protocol is not established.	Reviewed 2026-09-06
Terminal-Bench 2.0 (Terminus-2)	Coding	65.4% Reported settings & source Kimi K2.6 card, Terminal-Bench 2.0 (Terminus-2). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Coding: average of ten runs; Terminal uses Terminus2 preserve-thinking; SWE uses in-house SWE-agent. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, Terminal-Bench 2.0 (Terminus-2), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh contributes to capability	lab self-report	Reviewed 2026-09-12
Toolathlon	Agentic	54.6% Reported settings & source Kimi K2.6 card, Toolathlon. Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Cited from another official report; does not establish a new same-protocol evaluation. Comparison limit: Cited comparator requires original evaluation provenance; not a new matched run. Kimi K2.6 official model card · Evaluation Results table, Toolathlon, column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Cited comparator requires original evaluation provenance; not a new matched run.	Reviewed 2026-09-12
V* (w/ python)	Supporting evidence	98.4% Reported settings & source Kimi K2.6 card, V* (w/ python). Thinking mode; temperature 1.0, top-p 1.0, context 262144. Comparator settings: GPT-5.4 xhigh, Opus4.6 max, Gemini3.1Pro high. See benchmark-specific footnotes. Only own K2.6 and starred re-evaluations can form new comparisons. Vision: 98304 tokens, average of three runs; Python condition uses 65536 tokens per step, at most50steps. Source SHA256 95db3be1d0473e482c7aa901f237ad7af8bb7929625a777b27428081bb8ea9f7. Own measurement or explicitly starred same-condition re-evaluation. Assess benchmark-specific protocol before admission. Kimi K2.6 official model card · Evaluation Results table, V* (w/ python), column 3 · reviewed 2026-09-12	Published configuration Effort: XHigh	lab self-report Supporting evidence outside the reviewed capability core	Reviewed 2026-09-12
GDPval-AA	Agentic	—	—	—	—
GPQA Diamond	Hard reasoning	—	—	—	—
Humanity's Last Exam	Hard reasoning	—	—	—	—
LiveCodeBench	Coding	—	—	—	—
OSWorld-Verified	Agentic	—	—	—	—
SWE-bench Verified	Agentic	—	—	—	—
Terminal-Bench 2.1	Agentic	—	—	—	—